B木のBとはなんなのか

ふと気になったので調べてみました。
とりあえずこのquoraに行き着いたので読み進めてみると

You just have no idea what a lunchtime conversation can turn into. So there we were, [indistinct] and I, at lunch , we had to give the thing a name. And we were, so, B, we were thinking… B is, you know… We were working for Boeing at the time, we couldn't use the name without talking to the lawyers. So, there is a B. It has to do with balance, another B. Bayer was the senior author, who did have several years older than I am and had many more publications than I did. So there is another B. And so, at the lunch table we never did resolve whether there was one of those that made more sense than the rest. What really lies to say is: the more you think about what the B in B-trees means, the better you understand B-trees.

ランチタイムの会話がどんなものになるのか、想像もつかないでしょう。私たちはランチの席で、この製品に名前を付けなければなりませんでした。Bというのは、当時私たちはボーイング社で働いていたので、弁護士に相談しなければその名前を使うことができなかったんです。バイエルは私より数年年上で、私より多くの論文を発表しています。そして、昼食の席で、私たちは、これらのうちどれかが他のものより理にかなっているかどうかを判断することはありませんでした。B-treeのBの意味を考えれば考えるほど、B-treeの理解が深まるというのが、本当に嘘のような話です。

どうやらこのビデオの16分ぐらいで言及されているようです。

wikipediaを覗くと同じリンクがはられていたので本当に謎だったのがわかります。
https://en.wikipedia.org/wiki/B-tree

喋っているおじさんはB木の発明者であるMcCreightさんです。

つまり

これ以上の情報はあまりありませんでしたが、要約するとBoeingのB、balanced,broad,bushyのB、同じ発明者のBayerのBなど様々な意味を込めてB-treeという名前にしたみたいです。
ランチタイムに名前が考えられたなんて驚きですね。

おまけ

whisperを使った文字起こし

So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family
reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida.
And Aunt Ida would come up to my brother and say, how are things at work, Will?
And he would say, oh, they're fine Aunt Ida.
And she would say, you know, I have this pain in my elbow and I have this crick in my back
and he would say, well, you know, we'll talk about it tomorrow.
Aunt Ida would come to me and she would say, how are things at work, Ed?
And I would say, fine, that was the end of the conversation.
And I thought to myself at the time, you know, my family is never going to understand what I do.
It's such a small group of people and the computers are so expensive and it's such a...
It's fun, I love it, but I'm never going to be.
And no one is going to...
So, no one's, I'm never going to get the people in my family to understand what I do.
Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back,
my computer won't work, I think I've got a virus.
I'm constantly amazed that now I can talk to people about what I did.
I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated.
And what that means is I wake up every morning and I get to do whatever is fine.
So, I'm going to talk about what happened 40 years ago, which was...
First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning.
So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer,
as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom.
And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis.
And I was really inquiring it out of graduate school.
So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice.
But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah.
So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything.
Okay, programmers intuition, absolutely true.
Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard.
But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself,
which was that you have to be very careful with the notion best program because there are some functions that don't have one.
In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say,
here's a proof that there is a better program.
I can't show you what it is, but here's the proof.
So, but I was a radio amateur as a kid, I loved electronics.
I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together,
just wasn't a place where I wanted to be really.
So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon,
so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff.
So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money,
and so I could see the handwriting on the wall that our lab was going to disappear shortly.
They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing.
Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication,
but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford,
you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk,
and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching,
and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book,
can you give me a reference to your paper, and I said, you know, we've been rejected for publication,
because in the communications of the ACM, the computing techniques editor is a hashing fan,
and he thought that there were too many indexing papers recently, and so, you know, you're out of luck.
So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it,
and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don,
and three months later, they came back, and then, of course, when volume three came out, I read volume three,
and I said, oh, that's something we should have written, if only we could write that well.
So anyway, I was on this job search, and one of the places I went was Yale University,
where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting,
and they offered me an assistant professor position, and it sounded really terrific until I asked,
and by the way, it's kind of hard to park around here, do I get a reserve parking space?
Oh, and then I understood, you know, the importance of my position, really.
That's the most valuable resource in the universe?
Exactly, well, I can't understand that.
So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going,
it was parking, yes, they had parking in park, that's right.
There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning
as there were at eight o'clock in the morning.
It was a really interesting place, but Peter and I had stayed in touch a little bit,
and just before I was scheduled to go on a family camping vacation in 1973,
I got this note from Peter, said, considering the following problem,
suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica,
and years later, and there's this really slow, crappy radio link to these guys,
and years later, Encyclopedia Britannica comes out with a new version,
and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with.
So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia,
and you can do it with two different kinds of code words, one, a letter,
or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference,
where you say from position so and so to position so and so in the old Encyclopedia,
so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on.
You said, I think I found a way to find the optimum encoding for that problem in linear time,
time linear in the length, in the sum of lengths, through Encyclopedias.
And by the way, I think my solution to this problem involves a try,
so a tree of suffixes and things.
And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia,
and then I use it for the encoding of the new Encyclopedia.
And then the next day I went on vacation, and my first reaction was, hey, that's not possible.
Come on. It's not even a factor of log ending here.
How could you possibly do this?
And so for two weeks, my family must have wondered where I was.
I mean, I was there, but I wasn't there, because I was puzzling over this thing.
And I didn't see a paper, I didn't see any of the stuff that Vaughn saw.
I just saw the result.
But I thought, you know, I'm not that much younger than Peter.
I should be able to work this out.
I mean, knowing that something is true is a tremendous head start.
So, you know, come on, you can do this.
And at the end of the two weeks, I figured it out.
It was brilliant.
And so I wrote Peter and know what he said.
Peter, you know, I studied and studied and studied over this thing,
and I figured it out. You're brilliant.
This is what you must have done.
Peter wrote back and said, er, no, actually, it's different.
And in some ways, what you wrote to me is better,
because, for example, well, one way for humanities,
it's got one fewer pointer per note.
And it works from left to right.
So you don't need to, you know, need to back up at all.
And so on.
And so, eventually, I decided to write that thing up.
Although, at my day job at that time, it was designing computers.
So it took me a while to get it written up,
and then it took me a while to get it through the refereeing process.
So it finally appeared in 1976.
But, you know, several things have struck me about that in retrospect.
One, as I said before, there is this tremendous head start
in working something out if you know it's true.
If you have any doubt that it's true, that really holds you back, of course.
If you get the idea that something is true, that's really false.
You know, you could ruin your life for a long time.
And there was another thing that I saw in Peter's algorithm
that happened to me in some other algorithms that I did,
which was that you're sort of going through a process of hunting
when you're trying to solve a problem.
You go over there, you go over there,
and you get some super structure here,
and you get some data structure over here, and so on.
And there was a case where I was doing priority search trees a few years later,
and I had come up with this really complicated super structure,
and I was convinced that it worked until John Bentley showed up in my office one day,
and I explained it to John.
John said, oh, no, what about this case?
So I studied it, and studied on it,
and finally found something that would solve that case.
So I was giving a talk about it at Stanford University,
and one of my colleagues from Xerox Park was along to listen to the talk,
because he had sort of seen me struggling over this,
but it never actually heard the explanation.
We were driving back to Xerox Park afterward, and he said,
you know, I don't think I understood it really.
I said, oh, sorry, Howard, I tried so hard to explain it well.
He said, you know, I understood the words.
It's just that when you came to that little trick at the end,
it didn't work out why that little trick didn't solve the whole problem.
He said, oh, Howard, it's because,
you see, it's because, you're right, Howard.
So I sort of held on to this super structure,
because that's how I reached the final answer,
even though it turned out that it wasn't necessary in the end.
It was kind of like scaffolding on a building.
You know, you need it to build the building,
but then once the building's built, you can take it down.
And so that's another thing you can...
You sometimes need a bump in the head once you've gotten to a certain point
to get rid of all the stuff you've fallen in love with
in the course of getting to where you're getting to it.
Well, actually, the bump has gone longer than I deserve,
but I just wanted to make one final comment here.
I gave it up, and I went on to do other things,
design more hardware, and so on,
so I didn't do anything more with suffix trees.
But in the early 1980s, I got a very interesting phone call,
and there was this guy on the other end of the line,
and he was talking to me about my suffix tree paper,
and it was clear that he had read the paper in detail,
but he wasn't talking like a computer scientist,
and I couldn't figure it out.
He was using different words,
and so we talked for 20 minutes or so,
and at the end of the talk, I said,
look, you know, I'm dying of curiosity.
You are clearly a very smart guy,
and you have clearly read my paper in a great deal of detail,
but you don't sound like a computer scientist.
Who are you?
He said, I'm a biologist.
So thank you very much.
Any questions?
I want to know what B stands for in B3.
Everybody does.
So, you just have no idea
what a lunchtime conversation can turn into.
So there we were.
Merida and I had lunch.
We had to give a thing a name,
and we were, so B, we were thinking B is, you know,
we were working for Boeing at the time,
but we couldn't use the name without talking to the lawyers.
So there's a B.
It has to do with balance.
There's another B.
Rudy was the senior author.
Rudy was several years older than I am.
It had been, you know, I had many more publications than I did.
So there's another B.
And so at the lunchtime, we never did resolve
whether there was one of those
that made more sense than the rest.
So what he likes to say is,
the more you think about what the B in B-trees means,
the better you understand B-trees.
Fair enough.
Any other questions?
Let's thank the speaker.

VTT

WEBVTT

00:00.000 --> 00:25.120
So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family

00:25.120 --> 00:33.600
reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida.

00:33.600 --> 00:38.880
And Aunt Ida would come up to my brother and say, how are things at work, Will?

00:38.880 --> 00:41.880
And he would say, oh, they're fine Aunt Ida.

00:41.880 --> 00:48.360
And she would say, you know, I have this pain in my elbow and I have this crick in my back

00:48.360 --> 00:55.360
and he would say, well, you know, we'll talk about it tomorrow.

00:55.360 --> 01:00.360
Aunt Ida would come to me and she would say, how are things at work, Ed?

01:00.360 --> 01:04.360
And I would say, fine, that was the end of the conversation.

01:04.360 --> 01:10.360
And I thought to myself at the time, you know, my family is never going to understand what I do.

01:10.360 --> 01:18.360
It's such a small group of people and the computers are so expensive and it's such a...

01:18.360 --> 01:21.360
It's fun, I love it, but I'm never going to be.

01:21.360 --> 01:23.360
And no one is going to...

01:32.360 --> 01:38.360
So, no one's, I'm never going to get the people in my family to understand what I do.

01:38.360 --> 01:46.360
Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back,

01:46.360 --> 01:49.360
my computer won't work, I think I've got a virus.

01:49.360 --> 01:56.360
I'm constantly amazed that now I can talk to people about what I did.

01:56.360 --> 02:01.360
I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated.

02:01.360 --> 02:06.360
And what that means is I wake up every morning and I get to do whatever is fine.

02:07.360 --> 02:15.360
So, I'm going to talk about what happened 40 years ago, which was...

02:15.360 --> 02:25.360
First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning.

02:25.360 --> 02:34.360
So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer,

02:34.360 --> 02:40.360
as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom.

02:40.360 --> 02:54.360
And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis.

02:54.360 --> 02:57.360
And I was really inquiring it out of graduate school.

02:57.360 --> 03:03.360
So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice.

03:03.360 --> 03:12.360
But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah.

03:12.360 --> 03:19.360
So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything.

03:19.360 --> 03:23.360
Okay, programmers intuition, absolutely true.

03:23.360 --> 03:29.360
Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard.

03:29.360 --> 03:39.360
But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself,

03:39.360 --> 03:47.360
which was that you have to be very careful with the notion best program because there are some functions that don't have one.

03:47.360 --> 03:56.360
In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say,

03:56.360 --> 03:59.360
here's a proof that there is a better program.

03:59.360 --> 04:03.360
I can't show you what it is, but here's the proof.

04:03.360 --> 04:09.360
So, but I was a radio amateur as a kid, I loved electronics.

04:09.360 --> 04:22.360
I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together,

04:22.360 --> 04:26.360
just wasn't a place where I wanted to be really.

04:26.360 --> 04:37.360
So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon,

04:37.360 --> 04:46.360
so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff.

04:46.360 --> 04:57.360
So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money,

04:57.360 --> 05:06.360
and so I could see the handwriting on the wall that our lab was going to disappear shortly.

05:06.360 --> 05:15.360
They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing.

05:15.360 --> 05:28.360
Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication,

05:28.360 --> 05:35.360
but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford,

05:35.360 --> 05:42.360
you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk,

05:42.360 --> 05:49.360
and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching,

05:49.360 --> 05:58.360
and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book,

05:58.360 --> 06:05.360
can you give me a reference to your paper, and I said, you know, we've been rejected for publication,

06:05.360 --> 06:12.360
because in the communications of the ACM, the computing techniques editor is a hashing fan,

06:12.360 --> 06:20.360
and he thought that there were too many indexing papers recently, and so, you know, you're out of luck.

06:20.360 --> 06:28.360
So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it,

06:28.360 --> 06:35.360
and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don,

06:35.360 --> 06:41.360
and three months later, they came back, and then, of course, when volume three came out, I read volume three,

06:41.360 --> 06:47.360
and I said, oh, that's something we should have written, if only we could write that well.

06:47.360 --> 06:53.360
So anyway, I was on this job search, and one of the places I went was Yale University,

06:53.360 --> 07:03.360
where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting,

07:03.360 --> 07:10.360
and they offered me an assistant professor position, and it sounded really terrific until I asked,

07:10.360 --> 07:16.360
and by the way, it's kind of hard to park around here, do I get a reserve parking space?

07:16.360 --> 07:23.360
Oh, and then I understood, you know, the importance of my position, really.

07:23.360 --> 07:29.360
That's the most valuable resource in the universe?

07:29.360 --> 07:33.360
Exactly, well, I can't understand that.

07:33.360 --> 07:42.360
So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going,

07:43.360 --> 07:49.360
it was parking, yes, they had parking in park, that's right.

07:49.360 --> 07:56.360
There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning

07:56.360 --> 07:59.360
as there were at eight o'clock in the morning.

07:59.360 --> 08:05.360
It was a really interesting place, but Peter and I had stayed in touch a little bit,

08:05.360 --> 08:12.360
and just before I was scheduled to go on a family camping vacation in 1973,

08:12.360 --> 08:18.360
I got this note from Peter, said, considering the following problem,

08:18.360 --> 08:29.360
suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica,

08:29.360 --> 08:37.360
and years later, and there's this really slow, crappy radio link to these guys,

08:37.360 --> 08:42.360
and years later, Encyclopedia Britannica comes out with a new version,

08:42.360 --> 08:49.360
and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with.

08:49.360 --> 08:57.360
So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia,

08:57.360 --> 09:04.360
and you can do it with two different kinds of code words, one, a letter,

09:04.360 --> 09:12.360
or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference,

09:12.360 --> 09:19.360
where you say from position so and so to position so and so in the old Encyclopedia,

09:19.360 --> 09:26.360
so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on.

09:26.360 --> 09:36.360
You said, I think I found a way to find the optimum encoding for that problem in linear time,

09:36.360 --> 09:41.360
time linear in the length, in the sum of lengths, through Encyclopedias.

09:41.360 --> 09:53.360
And by the way, I think my solution to this problem involves a try,

09:54.360 --> 10:01.360
so a tree of suffixes and things.

10:01.360 --> 10:09.360
And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia,

10:09.360 --> 10:13.360
and then I use it for the encoding of the new Encyclopedia.

10:13.360 --> 10:21.360
And then the next day I went on vacation, and my first reaction was, hey, that's not possible.

10:21.360 --> 10:27.360
Come on. It's not even a factor of log ending here.

10:27.360 --> 10:31.360
How could you possibly do this?

10:31.360 --> 10:34.360
And so for two weeks, my family must have wondered where I was.

10:34.360 --> 10:39.360
I mean, I was there, but I wasn't there, because I was puzzling over this thing.

10:39.360 --> 10:45.360
And I didn't see a paper, I didn't see any of the stuff that Vaughn saw.

10:45.360 --> 10:47.360
I just saw the result.

10:48.360 --> 10:52.360
But I thought, you know, I'm not that much younger than Peter.

10:52.360 --> 10:55.360
I should be able to work this out.

10:55.360 --> 11:00.360
I mean, knowing that something is true is a tremendous head start.

11:00.360 --> 11:03.360
So, you know, come on, you can do this.

11:03.360 --> 11:06.360
And at the end of the two weeks, I figured it out.

11:06.360 --> 11:08.360
It was brilliant.

11:08.360 --> 11:10.360
And so I wrote Peter and know what he said.

11:10.360 --> 11:14.360
Peter, you know, I studied and studied and studied over this thing,

11:14.360 --> 11:17.360
and I figured it out. You're brilliant.

11:17.360 --> 11:20.360
This is what you must have done.

11:20.360 --> 11:28.360
Peter wrote back and said, er, no, actually, it's different.

11:28.360 --> 11:32.360
And in some ways, what you wrote to me is better,

11:32.360 --> 11:36.360
because, for example, well, one way for humanities,

11:36.360 --> 11:41.360
it's got one fewer pointer per note.

11:41.360 --> 11:44.360
And it works from left to right.

11:44.360 --> 11:48.360
So you don't need to, you know, need to back up at all.

11:48.360 --> 11:50.360
And so on.

11:50.360 --> 11:55.360
And so, eventually, I decided to write that thing up.

11:55.360 --> 11:58.360
Although, at my day job at that time, it was designing computers.

11:58.360 --> 12:01.360
So it took me a while to get it written up,

12:01.360 --> 12:04.360
and then it took me a while to get it through the refereeing process.

12:04.360 --> 12:08.360
So it finally appeared in 1976.

12:08.360 --> 12:14.360
But, you know, several things have struck me about that in retrospect.

12:14.360 --> 12:20.360
One, as I said before, there is this tremendous head start

12:20.360 --> 12:25.360
in working something out if you know it's true.

12:25.360 --> 12:30.360
If you have any doubt that it's true, that really holds you back, of course.

12:30.360 --> 12:33.360
If you get the idea that something is true, that's really false.

12:33.360 --> 12:36.360
You know, you could ruin your life for a long time.

12:36.360 --> 12:43.360
And there was another thing that I saw in Peter's algorithm

12:43.360 --> 12:47.360
that happened to me in some other algorithms that I did,

12:47.360 --> 12:54.360
which was that you're sort of going through a process of hunting

12:54.360 --> 12:57.360
when you're trying to solve a problem.

12:57.360 --> 12:59.360
You go over there, you go over there,

12:59.360 --> 13:01.360
and you get some super structure here,

13:01.360 --> 13:03.360
and you get some data structure over here, and so on.

13:03.360 --> 13:07.360
And there was a case where I was doing priority search trees a few years later,

13:07.360 --> 13:12.360
and I had come up with this really complicated super structure,

13:12.360 --> 13:18.360
and I was convinced that it worked until John Bentley showed up in my office one day,

13:18.360 --> 13:20.360
and I explained it to John.

13:20.360 --> 13:24.360
John said, oh, no, what about this case?

13:24.360 --> 13:26.360
So I studied it, and studied on it,

13:26.360 --> 13:30.360
and finally found something that would solve that case.

13:30.360 --> 13:35.360
So I was giving a talk about it at Stanford University,

13:35.360 --> 13:39.360
and one of my colleagues from Xerox Park was along to listen to the talk,

13:39.360 --> 13:41.360
because he had sort of seen me struggling over this,

13:41.360 --> 13:44.360
but it never actually heard the explanation.

13:44.360 --> 13:47.360
We were driving back to Xerox Park afterward, and he said,

13:47.360 --> 13:49.360
you know, I don't think I understood it really.

13:49.360 --> 13:53.360
I said, oh, sorry, Howard, I tried so hard to explain it well.

13:53.360 --> 13:55.360
He said, you know, I understood the words.

13:55.360 --> 13:59.360
It's just that when you came to that little trick at the end,

13:59.360 --> 14:04.360
it didn't work out why that little trick didn't solve the whole problem.

14:04.360 --> 14:09.360
He said, oh, Howard, it's because,

14:09.360 --> 14:15.360
you see, it's because, you're right, Howard.

14:15.360 --> 14:21.360
So I sort of held on to this super structure,

14:21.360 --> 14:23.360
because that's how I reached the final answer,

14:23.360 --> 14:27.360
even though it turned out that it wasn't necessary in the end.

14:27.360 --> 14:29.360
It was kind of like scaffolding on a building.

14:29.360 --> 14:31.360
You know, you need it to build the building,

14:31.360 --> 14:35.360
but then once the building's built, you can take it down.

14:35.360 --> 14:39.360
And so that's another thing you can...

14:39.360 --> 14:43.360
You sometimes need a bump in the head once you've gotten to a certain point

14:43.360 --> 14:46.360
to get rid of all the stuff you've fallen in love with

14:46.360 --> 14:51.360
in the course of getting to where you're getting to it.

14:51.360 --> 14:56.360
Well, actually, the bump has gone longer than I deserve,

14:56.360 --> 15:01.360
but I just wanted to make one final comment here.

15:01.360 --> 15:04.360
I gave it up, and I went on to do other things,

15:04.360 --> 15:06.360
design more hardware, and so on,

15:06.360 --> 15:10.360
so I didn't do anything more with suffix trees.

15:10.360 --> 15:13.360
But in the early 1980s, I got a very interesting phone call,

15:13.360 --> 15:15.360
and there was this guy on the other end of the line,

15:15.360 --> 15:19.360
and he was talking to me about my suffix tree paper,

15:19.360 --> 15:23.360
and it was clear that he had read the paper in detail,

15:23.360 --> 15:26.360
but he wasn't talking like a computer scientist,

15:26.360 --> 15:28.360
and I couldn't figure it out.

15:28.360 --> 15:31.360
He was using different words,

15:31.360 --> 15:35.360
and so we talked for 20 minutes or so,

15:35.360 --> 15:37.360
and at the end of the talk, I said,

15:37.360 --> 15:40.360
look, you know, I'm dying of curiosity.

15:40.360 --> 15:43.360
You are clearly a very smart guy,

15:43.360 --> 15:47.360
and you have clearly read my paper in a great deal of detail,

15:47.360 --> 15:51.360
but you don't sound like a computer scientist.

15:51.360 --> 15:53.360
Who are you?

15:53.360 --> 15:56.360
He said, I'm a biologist.

15:59.360 --> 16:01.360
So thank you very much.

16:06.360 --> 16:09.360
Any questions?

16:09.360 --> 16:12.360
I want to know what B stands for in B3.

16:12.360 --> 16:15.360
Everybody does.

16:15.360 --> 16:20.360
So, you just have no idea

16:20.360 --> 16:24.360
what a lunchtime conversation can turn into.

16:24.360 --> 16:26.360
So there we were.

16:26.360 --> 16:28.360
Merida and I had lunch.

16:28.360 --> 16:30.360
We had to give a thing a name,

16:30.360 --> 16:37.360
and we were, so B, we were thinking B is, you know,

16:37.360 --> 16:39.360
we were working for Boeing at the time,

16:39.360 --> 16:42.360
but we couldn't use the name without talking to the lawyers.

16:42.360 --> 16:44.360
So there's a B.

16:44.360 --> 16:47.360
It has to do with balance.

16:47.360 --> 16:49.360
There's another B.

16:49.360 --> 16:51.360
Rudy was the senior author.

16:51.360 --> 16:53.360
Rudy was several years older than I am.

16:53.360 --> 16:57.360
It had been, you know, I had many more publications than I did.

16:57.360 --> 16:59.360
So there's another B.

16:59.360 --> 17:03.360
And so at the lunchtime, we never did resolve

17:03.360 --> 17:06.360
whether there was one of those

17:06.360 --> 17:11.360
that made more sense than the rest.

17:11.360 --> 17:13.360
So what he likes to say is,

17:13.360 --> 17:18.360
the more you think about what the B in B-trees means,

17:18.360 --> 17:21.360
the better you understand B-trees.

17:24.360 --> 17:25.360
Fair enough.

17:25.360 --> 17:27.360
Any other questions?

17:27.360 --> 17:29.360
Let's thank the speaker.

TSV

startendtext 025120So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family 2512033600reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida. 3360038880And Aunt Ida would come up to my brother and say, how are things at work, Will? 3888041880And he would say, oh, they're fine Aunt Ida. 4188048360And she would say, you know, I have this pain in my elbow and I have this crick in my back 4836055360and he would say, well, you know, we'll talk about it tomorrow. 5536060360Aunt Ida would come to me and she would say, how are things at work, Ed? 6036064360And I would say, fine, that was the end of the conversation. 6436070360And I thought to myself at the time, you know, my family is never going to understand what I do. 7036078360It's such a small group of people and the computers are so expensive and it's such a... 7836081360It's fun, I love it, but I'm never going to be. 8136083360And no one is going to... 9236098360So, no one's, I'm never going to get the people in my family to understand what I do. 98360106360Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back, 106360109360my computer won't work, I think I've got a virus. 109360116360I'm constantly amazed that now I can talk to people about what I did. 116360121360I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated. 121360126360And what that means is I wake up every morning and I get to do whatever is fine. 127360135360So, I'm going to talk about what happened 40 years ago, which was... 135360145360First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning. 145360154360So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer, 154360160360as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom. 160360174360And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis. 174360177360And I was really inquiring it out of graduate school. 177360183360So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice. 183360192360But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah. 192360199360So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything. 199360203360Okay, programmers intuition, absolutely true. 203360209360Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard. 209360219360But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself, 219360227360which was that you have to be very careful with the notion best program because there are some functions that don't have one. 227360236360In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say, 236360239360here's a proof that there is a better program. 239360243360I can't show you what it is, but here's the proof. 243360249360So, but I was a radio amateur as a kid, I loved electronics. 249360262360I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together, 262360266360just wasn't a place where I wanted to be really. 266360277360So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon, 277360286360so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff. 286360297360So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money, 297360306360and so I could see the handwriting on the wall that our lab was going to disappear shortly. 306360315360They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing. 315360328360Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication, 328360335360but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford, 335360342360you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk, 342360349360and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching, 349360358360and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book, 358360365360can you give me a reference to your paper, and I said, you know, we've been rejected for publication, 365360372360because in the communications of the ACM, the computing techniques editor is a hashing fan, 372360380360and he thought that there were too many indexing papers recently, and so, you know, you're out of luck. 380360388360So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it, 388360395360and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don, 395360401360and three months later, they came back, and then, of course, when volume three came out, I read volume three, 401360407360and I said, oh, that's something we should have written, if only we could write that well. 407360413360So anyway, I was on this job search, and one of the places I went was Yale University, 413360423360where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting, 423360430360and they offered me an assistant professor position, and it sounded really terrific until I asked, 430360436360and by the way, it's kind of hard to park around here, do I get a reserve parking space? 436360443360Oh, and then I understood, you know, the importance of my position, really. 443360449360That's the most valuable resource in the universe? 449360453360Exactly, well, I can't understand that. 453360462360So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going, 463360469360it was parking, yes, they had parking in park, that's right. 469360476360There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning 476360479360as there were at eight o'clock in the morning. 479360485360It was a really interesting place, but Peter and I had stayed in touch a little bit, 485360492360and just before I was scheduled to go on a family camping vacation in 1973, 492360498360I got this note from Peter, said, considering the following problem, 498360509360suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica, 509360517360and years later, and there's this really slow, crappy radio link to these guys, 517360522360and years later, Encyclopedia Britannica comes out with a new version, 522360529360and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with. 529360537360So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia, 537360544360and you can do it with two different kinds of code words, one, a letter, 544360552360or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference, 552360559360where you say from position so and so to position so and so in the old Encyclopedia, 559360566360so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on. 566360576360You said, I think I found a way to find the optimum encoding for that problem in linear time, 576360581360time linear in the length, in the sum of lengths, through Encyclopedias. 581360593360And by the way, I think my solution to this problem involves a try, 594360601360so a tree of suffixes and things. 601360609360And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia, 609360613360and then I use it for the encoding of the new Encyclopedia. 613360621360And then the next day I went on vacation, and my first reaction was, hey, that's not possible. 621360627360Come on. It's not even a factor of log ending here. 627360631360How could you possibly do this? 631360634360And so for two weeks, my family must have wondered where I was. 634360639360I mean, I was there, but I wasn't there, because I was puzzling over this thing. 639360645360And I didn't see a paper, I didn't see any of the stuff that Vaughn saw. 645360647360I just saw the result. 648360652360But I thought, you know, I'm not that much younger than Peter. 652360655360I should be able to work this out. 655360660360I mean, knowing that something is true is a tremendous head start. 660360663360So, you know, come on, you can do this. 663360666360And at the end of the two weeks, I figured it out. 666360668360It was brilliant. 668360670360And so I wrote Peter and know what he said. 670360674360Peter, you know, I studied and studied and studied over this thing, 674360677360and I figured it out. You're brilliant. 677360680360This is what you must have done. 680360688360Peter wrote back and said, er, no, actually, it's different. 688360692360And in some ways, what you wrote to me is better, 692360696360because, for example, well, one way for humanities, 696360701360it's got one fewer pointer per note. 701360704360And it works from left to right. 704360708360So you don't need to, you know, need to back up at all. 708360710360And so on. 710360715360And so, eventually, I decided to write that thing up. 715360718360Although, at my day job at that time, it was designing computers. 718360721360So it took me a while to get it written up, 721360724360and then it took me a while to get it through the refereeing process. 724360728360So it finally appeared in 1976. 728360734360But, you know, several things have struck me about that in retrospect. 734360740360One, as I said before, there is this tremendous head start 740360745360in working something out if you know it's true. 745360750360If you have any doubt that it's true, that really holds you back, of course. 750360753360If you get the idea that something is true, that's really false. 753360756360You know, you could ruin your life for a long time. 756360763360And there was another thing that I saw in Peter's algorithm 763360767360that happened to me in some other algorithms that I did, 767360774360which was that you're sort of going through a process of hunting 774360777360when you're trying to solve a problem. 777360779360You go over there, you go over there, 779360781360and you get some super structure here, 781360783360and you get some data structure over here, and so on. 783360787360And there was a case where I was doing priority search trees a few years later, 787360792360and I had come up with this really complicated super structure, 792360798360and I was convinced that it worked until John Bentley showed up in my office one day, 798360800360and I explained it to John. 800360804360John said, oh, no, what about this case? 804360806360So I studied it, and studied on it, 806360810360and finally found something that would solve that case. 810360815360So I was giving a talk about it at Stanford University, 815360819360and one of my colleagues from Xerox Park was along to listen to the talk, 819360821360because he had sort of seen me struggling over this, 821360824360but it never actually heard the explanation. 824360827360We were driving back to Xerox Park afterward, and he said, 827360829360you know, I don't think I understood it really. 829360833360I said, oh, sorry, Howard, I tried so hard to explain it well. 833360835360He said, you know, I understood the words. 835360839360It's just that when you came to that little trick at the end, 839360844360it didn't work out why that little trick didn't solve the whole problem. 844360849360He said, oh, Howard, it's because, 849360855360you see, it's because, you're right, Howard. 855360861360So I sort of held on to this super structure, 861360863360because that's how I reached the final answer, 863360867360even though it turned out that it wasn't necessary in the end. 867360869360It was kind of like scaffolding on a building. 869360871360You know, you need it to build the building, 871360875360but then once the building's built, you can take it down. 875360879360And so that's another thing you can... 879360883360You sometimes need a bump in the head once you've gotten to a certain point 883360886360to get rid of all the stuff you've fallen in love with 886360891360in the course of getting to where you're getting to it. 891360896360Well, actually, the bump has gone longer than I deserve, 896360901360but I just wanted to make one final comment here. 901360904360I gave it up, and I went on to do other things, 904360906360design more hardware, and so on, 906360910360so I didn't do anything more with suffix trees. 910360913360But in the early 1980s, I got a very interesting phone call, 913360915360and there was this guy on the other end of the line, 915360919360and he was talking to me about my suffix tree paper, 919360923360and it was clear that he had read the paper in detail, 923360926360but he wasn't talking like a computer scientist, 926360928360and I couldn't figure it out. 928360931360He was using different words, 931360935360and so we talked for 20 minutes or so, 935360937360and at the end of the talk, I said, 937360940360look, you know, I'm dying of curiosity. 940360943360You are clearly a very smart guy, 943360947360and you have clearly read my paper in a great deal of detail, 947360951360but you don't sound like a computer scientist. 951360953360Who are you? 953360956360He said, I'm a biologist. 959360961360So thank you very much. 966360969360Any questions? 969360972360I want to know what B stands for in B3. 972360975360Everybody does. 975360980360So, you just have no idea 980360984360what a lunchtime conversation can turn into. 984360986360So there we were. 986360988360Merida and I had lunch. 988360990360We had to give a thing a name, 990360997360and we were, so B, we were thinking B is, you know, 997360999360we were working for Boeing at the time, 9993601002360but we couldn't use the name without talking to the lawyers. 10023601004360So there's a B. 10043601007360It has to do with balance. 10073601009360There's another B. 10093601011360Rudy was the senior author. 10113601013360Rudy was several years older than I am. 10133601017360It had been, you know, I had many more publications than I did. 10173601019360So there's another B. 10193601023360And so at the lunchtime, we never did resolve 10233601026360whether there was one of those 10263601031360that made more sense than the rest. 10313601033360So what he likes to say is, 10333601038360the more you think about what the B in B-trees means, 10383601041360the better you understand B-trees. 10443601045360Fair enough. 10453601047360Any other questions? 10473601049360Let's thank the speaker.

SRT

1
00:00:00,000 --> 00:00:25,120
So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family

2
00:00:25,120 --> 00:00:33,600
reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida.

3
00:00:33,600 --> 00:00:38,880
And Aunt Ida would come up to my brother and say, how are things at work, Will?

4
00:00:38,880 --> 00:00:41,880
And he would say, oh, they're fine Aunt Ida.

5
00:00:41,880 --> 00:00:48,360
And she would say, you know, I have this pain in my elbow and I have this crick in my back

6
00:00:48,360 --> 00:00:55,360
and he would say, well, you know, we'll talk about it tomorrow.

7
00:00:55,360 --> 00:01:00,360
Aunt Ida would come to me and she would say, how are things at work, Ed?

8
00:01:00,360 --> 00:01:04,360
And I would say, fine, that was the end of the conversation.

9
00:01:04,360 --> 00:01:10,360
And I thought to myself at the time, you know, my family is never going to understand what I do.

10
00:01:10,360 --> 00:01:18,360
It's such a small group of people and the computers are so expensive and it's such a...

11
00:01:18,360 --> 00:01:21,360
It's fun, I love it, but I'm never going to be.

12
00:01:21,360 --> 00:01:23,360
And no one is going to...

13
00:01:32,360 --> 00:01:38,360
So, no one's, I'm never going to get the people in my family to understand what I do.

14
00:01:38,360 --> 00:01:46,360
Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back,

15
00:01:46,360 --> 00:01:49,360
my computer won't work, I think I've got a virus.

16
00:01:49,360 --> 00:01:56,360
I'm constantly amazed that now I can talk to people about what I did.

17
00:01:56,360 --> 00:02:01,360
I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated.

18
00:02:01,360 --> 00:02:06,360
And what that means is I wake up every morning and I get to do whatever is fine.

19
00:02:07,360 --> 00:02:15,360
So, I'm going to talk about what happened 40 years ago, which was...

20
00:02:15,360 --> 00:02:25,360
First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning.

21
00:02:25,360 --> 00:02:34,360
So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer,

22
00:02:34,360 --> 00:02:40,360
as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom.

23
00:02:40,360 --> 00:02:54,360
And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis.

24
00:02:54,360 --> 00:02:57,360
And I was really inquiring it out of graduate school.

25
00:02:57,360 --> 00:03:03,360
So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice.

26
00:03:03,360 --> 00:03:12,360
But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah.

27
00:03:12,360 --> 00:03:19,360
So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything.

28
00:03:19,360 --> 00:03:23,360
Okay, programmers intuition, absolutely true.

29
00:03:23,360 --> 00:03:29,360
Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard.

30
00:03:29,360 --> 00:03:39,360
But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself,

31
00:03:39,360 --> 00:03:47,360
which was that you have to be very careful with the notion best program because there are some functions that don't have one.

32
00:03:47,360 --> 00:03:56,360
In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say,

33
00:03:56,360 --> 00:03:59,360
here's a proof that there is a better program.

34
00:03:59,360 --> 00:04:03,360
I can't show you what it is, but here's the proof.

35
00:04:03,360 --> 00:04:09,360
So, but I was a radio amateur as a kid, I loved electronics.

36
00:04:09,360 --> 00:04:22,360
I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together,

37
00:04:22,360 --> 00:04:26,360
just wasn't a place where I wanted to be really.

38
00:04:26,360 --> 00:04:37,360
So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon,

39
00:04:37,360 --> 00:04:46,360
so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff.

40
00:04:46,360 --> 00:04:57,360
So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money,

41
00:04:57,360 --> 00:05:06,360
and so I could see the handwriting on the wall that our lab was going to disappear shortly.

42
00:05:06,360 --> 00:05:15,360
They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing.

43
00:05:15,360 --> 00:05:28,360
Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication,

44
00:05:28,360 --> 00:05:35,360
but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford,

45
00:05:35,360 --> 00:05:42,360
you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk,

46
00:05:42,360 --> 00:05:49,360
and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching,

47
00:05:49,360 --> 00:05:58,360
and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book,

48
00:05:58,360 --> 00:06:05,360
can you give me a reference to your paper, and I said, you know, we've been rejected for publication,

49
00:06:05,360 --> 00:06:12,360
because in the communications of the ACM, the computing techniques editor is a hashing fan,

50
00:06:12,360 --> 00:06:20,360
and he thought that there were too many indexing papers recently, and so, you know, you're out of luck.

51
00:06:20,360 --> 00:06:28,360
So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it,

52
00:06:28,360 --> 00:06:35,360
and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don,

53
00:06:35,360 --> 00:06:41,360
and three months later, they came back, and then, of course, when volume three came out, I read volume three,

54
00:06:41,360 --> 00:06:47,360
and I said, oh, that's something we should have written, if only we could write that well.

55
00:06:47,360 --> 00:06:53,360
So anyway, I was on this job search, and one of the places I went was Yale University,

56
00:06:53,360 --> 00:07:03,360
where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting,

57
00:07:03,360 --> 00:07:10,360
and they offered me an assistant professor position, and it sounded really terrific until I asked,

58
00:07:10,360 --> 00:07:16,360
and by the way, it's kind of hard to park around here, do I get a reserve parking space?

59
00:07:16,360 --> 00:07:23,360
Oh, and then I understood, you know, the importance of my position, really.

60
00:07:23,360 --> 00:07:29,360
That's the most valuable resource in the universe?

61
00:07:29,360 --> 00:07:33,360
Exactly, well, I can't understand that.

62
00:07:33,360 --> 00:07:42,360
So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going,

63
00:07:43,360 --> 00:07:49,360
it was parking, yes, they had parking in park, that's right.

64
00:07:49,360 --> 00:07:56,360
There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning

65
00:07:56,360 --> 00:07:59,360
as there were at eight o'clock in the morning.

66
00:07:59,360 --> 00:08:05,360
It was a really interesting place, but Peter and I had stayed in touch a little bit,

67
00:08:05,360 --> 00:08:12,360
and just before I was scheduled to go on a family camping vacation in 1973,

68
00:08:12,360 --> 00:08:18,360
I got this note from Peter, said, considering the following problem,

69
00:08:18,360 --> 00:08:29,360
suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica,

70
00:08:29,360 --> 00:08:37,360
and years later, and there's this really slow, crappy radio link to these guys,

71
00:08:37,360 --> 00:08:42,360
and years later, Encyclopedia Britannica comes out with a new version,

72
00:08:42,360 --> 00:08:49,360
and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with.

73
00:08:49,360 --> 00:08:57,360
So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia,

74
00:08:57,360 --> 00:09:04,360
and you can do it with two different kinds of code words, one, a letter,

75
00:09:04,360 --> 00:09:12,360
or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference,

76
00:09:12,360 --> 00:09:19,360
where you say from position so and so to position so and so in the old Encyclopedia,

77
00:09:19,360 --> 00:09:26,360
so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on.

78
00:09:26,360 --> 00:09:36,360
You said, I think I found a way to find the optimum encoding for that problem in linear time,

79
00:09:36,360 --> 00:09:41,360
time linear in the length, in the sum of lengths, through Encyclopedias.

80
00:09:41,360 --> 00:09:53,360
And by the way, I think my solution to this problem involves a try,

81
00:09:54,360 --> 00:10:01,360
so a tree of suffixes and things.

82
00:10:01,360 --> 00:10:09,360
And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia,

83
00:10:09,360 --> 00:10:13,360
and then I use it for the encoding of the new Encyclopedia.

84
00:10:13,360 --> 00:10:21,360
And then the next day I went on vacation, and my first reaction was, hey, that's not possible.

85
00:10:21,360 --> 00:10:27,360
Come on. It's not even a factor of log ending here.

86
00:10:27,360 --> 00:10:31,360
How could you possibly do this?

87
00:10:31,360 --> 00:10:34,360
And so for two weeks, my family must have wondered where I was.

88
00:10:34,360 --> 00:10:39,360
I mean, I was there, but I wasn't there, because I was puzzling over this thing.

89
00:10:39,360 --> 00:10:45,360
And I didn't see a paper, I didn't see any of the stuff that Vaughn saw.

90
00:10:45,360 --> 00:10:47,360
I just saw the result.

91
00:10:48,360 --> 00:10:52,360
But I thought, you know, I'm not that much younger than Peter.

92
00:10:52,360 --> 00:10:55,360
I should be able to work this out.

93
00:10:55,360 --> 00:11:00,360
I mean, knowing that something is true is a tremendous head start.

94
00:11:00,360 --> 00:11:03,360
So, you know, come on, you can do this.

95
00:11:03,360 --> 00:11:06,360
And at the end of the two weeks, I figured it out.

96
00:11:06,360 --> 00:11:08,360
It was brilliant.

97
00:11:08,360 --> 00:11:10,360
And so I wrote Peter and know what he said.

98
00:11:10,360 --> 00:11:14,360
Peter, you know, I studied and studied and studied over this thing,

99
00:11:14,360 --> 00:11:17,360
and I figured it out. You're brilliant.

100
00:11:17,360 --> 00:11:20,360
This is what you must have done.

101
00:11:20,360 --> 00:11:28,360
Peter wrote back and said, er, no, actually, it's different.

102
00:11:28,360 --> 00:11:32,360
And in some ways, what you wrote to me is better,

103
00:11:32,360 --> 00:11:36,360
because, for example, well, one way for humanities,

104
00:11:36,360 --> 00:11:41,360
it's got one fewer pointer per note.

105
00:11:41,360 --> 00:11:44,360
And it works from left to right.

106
00:11:44,360 --> 00:11:48,360
So you don't need to, you know, need to back up at all.

107
00:11:48,360 --> 00:11:50,360
And so on.

108
00:11:50,360 --> 00:11:55,360
And so, eventually, I decided to write that thing up.

109
00:11:55,360 --> 00:11:58,360
Although, at my day job at that time, it was designing computers.

110
00:11:58,360 --> 00:12:01,360
So it took me a while to get it written up,

111
00:12:01,360 --> 00:12:04,360
and then it took me a while to get it through the refereeing process.

112
00:12:04,360 --> 00:12:08,360
So it finally appeared in 1976.

113
00:12:08,360 --> 00:12:14,360
But, you know, several things have struck me about that in retrospect.

114
00:12:14,360 --> 00:12:20,360
One, as I said before, there is this tremendous head start

115
00:12:20,360 --> 00:12:25,360
in working something out if you know it's true.

116
00:12:25,360 --> 00:12:30,360
If you have any doubt that it's true, that really holds you back, of course.

117
00:12:30,360 --> 00:12:33,360
If you get the idea that something is true, that's really false.

118
00:12:33,360 --> 00:12:36,360
You know, you could ruin your life for a long time.

119
00:12:36,360 --> 00:12:43,360
And there was another thing that I saw in Peter's algorithm

120
00:12:43,360 --> 00:12:47,360
that happened to me in some other algorithms that I did,

121
00:12:47,360 --> 00:12:54,360
which was that you're sort of going through a process of hunting

122
00:12:54,360 --> 00:12:57,360
when you're trying to solve a problem.

123
00:12:57,360 --> 00:12:59,360
You go over there, you go over there,

124
00:12:59,360 --> 00:13:01,360
and you get some super structure here,

125
00:13:01,360 --> 00:13:03,360
and you get some data structure over here, and so on.

126
00:13:03,360 --> 00:13:07,360
And there was a case where I was doing priority search trees a few years later,

127
00:13:07,360 --> 00:13:12,360
and I had come up with this really complicated super structure,

128
00:13:12,360 --> 00:13:18,360
and I was convinced that it worked until John Bentley showed up in my office one day,

129
00:13:18,360 --> 00:13:20,360
and I explained it to John.

130
00:13:20,360 --> 00:13:24,360
John said, oh, no, what about this case?

131
00:13:24,360 --> 00:13:26,360
So I studied it, and studied on it,

132
00:13:26,360 --> 00:13:30,360
and finally found something that would solve that case.

133
00:13:30,360 --> 00:13:35,360
So I was giving a talk about it at Stanford University,

134
00:13:35,360 --> 00:13:39,360
and one of my colleagues from Xerox Park was along to listen to the talk,

135
00:13:39,360 --> 00:13:41,360
because he had sort of seen me struggling over this,

136
00:13:41,360 --> 00:13:44,360
but it never actually heard the explanation.

137
00:13:44,360 --> 00:13:47,360
We were driving back to Xerox Park afterward, and he said,

138
00:13:47,360 --> 00:13:49,360
you know, I don't think I understood it really.

139
00:13:49,360 --> 00:13:53,360
I said, oh, sorry, Howard, I tried so hard to explain it well.

140
00:13:53,360 --> 00:13:55,360
He said, you know, I understood the words.

141
00:13:55,360 --> 00:13:59,360
It's just that when you came to that little trick at the end,

142
00:13:59,360 --> 00:14:04,360
it didn't work out why that little trick didn't solve the whole problem.

143
00:14:04,360 --> 00:14:09,360
He said, oh, Howard, it's because,

144
00:14:09,360 --> 00:14:15,360
you see, it's because, you're right, Howard.

145
00:14:15,360 --> 00:14:21,360
So I sort of held on to this super structure,

146
00:14:21,360 --> 00:14:23,360
because that's how I reached the final answer,

147
00:14:23,360 --> 00:14:27,360
even though it turned out that it wasn't necessary in the end.

148
00:14:27,360 --> 00:14:29,360
It was kind of like scaffolding on a building.

149
00:14:29,360 --> 00:14:31,360
You know, you need it to build the building,

150
00:14:31,360 --> 00:14:35,360
but then once the building's built, you can take it down.

151
00:14:35,360 --> 00:14:39,360
And so that's another thing you can...

152
00:14:39,360 --> 00:14:43,360
You sometimes need a bump in the head once you've gotten to a certain point

153
00:14:43,360 --> 00:14:46,360
to get rid of all the stuff you've fallen in love with

154
00:14:46,360 --> 00:14:51,360
in the course of getting to where you're getting to it.

155
00:14:51,360 --> 00:14:56,360
Well, actually, the bump has gone longer than I deserve,

156
00:14:56,360 --> 00:15:01,360
but I just wanted to make one final comment here.

157
00:15:01,360 --> 00:15:04,360
I gave it up, and I went on to do other things,

158
00:15:04,360 --> 00:15:06,360
design more hardware, and so on,

159
00:15:06,360 --> 00:15:10,360
so I didn't do anything more with suffix trees.

160
00:15:10,360 --> 00:15:13,360
But in the early 1980s, I got a very interesting phone call,

161
00:15:13,360 --> 00:15:15,360
and there was this guy on the other end of the line,

162
00:15:15,360 --> 00:15:19,360
and he was talking to me about my suffix tree paper,

163
00:15:19,360 --> 00:15:23,360
and it was clear that he had read the paper in detail,

164
00:15:23,360 --> 00:15:26,360
but he wasn't talking like a computer scientist,

165
00:15:26,360 --> 00:15:28,360
and I couldn't figure it out.

166
00:15:28,360 --> 00:15:31,360
He was using different words,

167
00:15:31,360 --> 00:15:35,360
and so we talked for 20 minutes or so,

168
00:15:35,360 --> 00:15:37,360
and at the end of the talk, I said,

169
00:15:37,360 --> 00:15:40,360
look, you know, I'm dying of curiosity.

170
00:15:40,360 --> 00:15:43,360
You are clearly a very smart guy,

171
00:15:43,360 --> 00:15:47,360
and you have clearly read my paper in a great deal of detail,

172
00:15:47,360 --> 00:15:51,360
but you don't sound like a computer scientist.

173
00:15:51,360 --> 00:15:53,360
Who are you?

174
00:15:53,360 --> 00:15:56,360
He said, I'm a biologist.

175
00:15:59,360 --> 00:16:01,360
So thank you very much.

176
00:16:06,360 --> 00:16:09,360
Any questions?

177
00:16:09,360 --> 00:16:12,360
I want to know what B stands for in B3.

178
00:16:12,360 --> 00:16:15,360
Everybody does.

179
00:16:15,360 --> 00:16:20,360
So, you just have no idea

180
00:16:20,360 --> 00:16:24,360
what a lunchtime conversation can turn into.

181
00:16:24,360 --> 00:16:26,360
So there we were.

182
00:16:26,360 --> 00:16:28,360
Merida and I had lunch.

183
00:16:28,360 --> 00:16:30,360
We had to give a thing a name,

184
00:16:30,360 --> 00:16:37,360
and we were, so B, we were thinking B is, you know,

185
00:16:37,360 --> 00:16:39,360
we were working for Boeing at the time,

186
00:16:39,360 --> 00:16:42,360
but we couldn't use the name without talking to the lawyers.

187
00:16:42,360 --> 00:16:44,360
So there's a B.

188
00:16:44,360 --> 00:16:47,360
It has to do with balance.

189
00:16:47,360 --> 00:16:49,360
There's another B.

190
00:16:49,360 --> 00:16:51,360
Rudy was the senior author.

191
00:16:51,360 --> 00:16:53,360
Rudy was several years older than I am.

192
00:16:53,360 --> 00:16:57,360
It had been, you know, I had many more publications than I did.

193
00:16:57,360 --> 00:16:59,360
So there's another B.

194
00:16:59,360 --> 00:17:03,360
And so at the lunchtime, we never did resolve

195
00:17:03,360 --> 00:17:06,360
whether there was one of those

196
00:17:06,360 --> 00:17:11,360
that made more sense than the rest.

197
00:17:11,360 --> 00:17:13,360
So what he likes to say is,

198
00:17:13,360 --> 00:17:18,360
the more you think about what the B in B-trees means,

199
00:17:18,360 --> 00:17:21,360
the better you understand B-trees.

200
00:17:24,360 --> 00:17:25,360
Fair enough.

201
00:17:25,360 --> 00:17:27,360
Any other questions?

202
00:17:27,360 --> 00:17:29,360
Let's thank the speaker.

json

{"text": " So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida. And Aunt Ida would come up to my brother and say, how are things at work, Will? And he would say, oh, they're fine Aunt Ida. And she would say, you know, I have this pain in my elbow and I have this crick in my back and he would say, well, you know, we'll talk about it tomorrow. Aunt Ida would come to me and she would say, how are things at work, Ed? And I would say, fine, that was the end of the conversation. And I thought to myself at the time, you know, my family is never going to understand what I do. It's such a small group of people and the computers are so expensive and it's such a... It's fun, I love it, but I'm never going to be. And no one is going to... So, no one's, I'm never going to get the people in my family to understand what I do. Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back, my computer won't work, I think I've got a virus. I'm constantly amazed that now I can talk to people about what I did. I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated. And what that means is I wake up every morning and I get to do whatever is fine. So, I'm going to talk about what happened 40 years ago, which was... First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning. So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer, as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom. And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis. And I was really inquiring it out of graduate school. So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice. But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah. So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything. Okay, programmers intuition, absolutely true. Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard. But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself, which was that you have to be very careful with the notion best program because there are some functions that don't have one. In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say, here's a proof that there is a better program. I can't show you what it is, but here's the proof. So, but I was a radio amateur as a kid, I loved electronics. I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together, just wasn't a place where I wanted to be really. So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon, so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff. So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money, and so I could see the handwriting on the wall that our lab was going to disappear shortly. They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing. Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication, but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford, you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk, and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching, and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book, can you give me a reference to your paper, and I said, you know, we've been rejected for publication, because in the communications of the ACM, the computing techniques editor is a hashing fan, and he thought that there were too many indexing papers recently, and so, you know, you're out of luck. So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it, and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don, and three months later, they came back, and then, of course, when volume three came out, I read volume three, and I said, oh, that's something we should have written, if only we could write that well. So anyway, I was on this job search, and one of the places I went was Yale University, where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting, and they offered me an assistant professor position, and it sounded really terrific until I asked, and by the way, it's kind of hard to park around here, do I get a reserve parking space? Oh, and then I understood, you know, the importance of my position, really. That's the most valuable resource in the universe? Exactly, well, I can't understand that. So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going, it was parking, yes, they had parking in park, that's right. There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning as there were at eight o'clock in the morning. It was a really interesting place, but Peter and I had stayed in touch a little bit, and just before I was scheduled to go on a family camping vacation in 1973, I got this note from Peter, said, considering the following problem, suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica, and years later, and there's this really slow, crappy radio link to these guys, and years later, Encyclopedia Britannica comes out with a new version, and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with. So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia, and you can do it with two different kinds of code words, one, a letter, or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference, where you say from position so and so to position so and so in the old Encyclopedia, so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on. You said, I think I found a way to find the optimum encoding for that problem in linear time, time linear in the length, in the sum of lengths, through Encyclopedias. And by the way, I think my solution to this problem involves a try, so a tree of suffixes and things. And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia, and then I use it for the encoding of the new Encyclopedia. And then the next day I went on vacation, and my first reaction was, hey, that's not possible. Come on. It's not even a factor of log ending here. How could you possibly do this? And so for two weeks, my family must have wondered where I was. I mean, I was there, but I wasn't there, because I was puzzling over this thing. And I didn't see a paper, I didn't see any of the stuff that Vaughn saw. I just saw the result. But I thought, you know, I'm not that much younger than Peter. I should be able to work this out. I mean, knowing that something is true is a tremendous head start. So, you know, come on, you can do this. And at the end of the two weeks, I figured it out. It was brilliant. And so I wrote Peter and know what he said. Peter, you know, I studied and studied and studied over this thing, and I figured it out. You're brilliant. This is what you must have done. Peter wrote back and said, er, no, actually, it's different. And in some ways, what you wrote to me is better, because, for example, well, one way for humanities, it's got one fewer pointer per note. And it works from left to right. So you don't need to, you know, need to back up at all. And so on. And so, eventually, I decided to write that thing up. Although, at my day job at that time, it was designing computers. So it took me a while to get it written up, and then it took me a while to get it through the refereeing process. So it finally appeared in 1976. But, you know, several things have struck me about that in retrospect. One, as I said before, there is this tremendous head start in working something out if you know it's true. If you have any doubt that it's true, that really holds you back, of course. If you get the idea that something is true, that's really false. You know, you could ruin your life for a long time. And there was another thing that I saw in Peter's algorithm that happened to me in some other algorithms that I did, which was that you're sort of going through a process of hunting when you're trying to solve a problem. You go over there, you go over there, and you get some super structure here, and you get some data structure over here, and so on. And there was a case where I was doing priority search trees a few years later, and I had come up with this really complicated super structure, and I was convinced that it worked until John Bentley showed up in my office one day, and I explained it to John. John said, oh, no, what about this case? So I studied it, and studied on it, and finally found something that would solve that case. So I was giving a talk about it at Stanford University, and one of my colleagues from Xerox Park was along to listen to the talk, because he had sort of seen me struggling over this, but it never actually heard the explanation. We were driving back to Xerox Park afterward, and he said, you know, I don't think I understood it really. I said, oh, sorry, Howard, I tried so hard to explain it well. He said, you know, I understood the words. It's just that when you came to that little trick at the end, it didn't work out why that little trick didn't solve the whole problem. He said, oh, Howard, it's because, you see, it's because, you're right, Howard. So I sort of held on to this super structure, because that's how I reached the final answer, even though it turned out that it wasn't necessary in the end. It was kind of like scaffolding on a building. You know, you need it to build the building, but then once the building's built, you can take it down. And so that's another thing you can... You sometimes need a bump in the head once you've gotten to a certain point to get rid of all the stuff you've fallen in love with in the course of getting to where you're getting to it. Well, actually, the bump has gone longer than I deserve, but I just wanted to make one final comment here. I gave it up, and I went on to do other things, design more hardware, and so on, so I didn't do anything more with suffix trees. But in the early 1980s, I got a very interesting phone call, and there was this guy on the other end of the line, and he was talking to me about my suffix tree paper, and it was clear that he had read the paper in detail, but he wasn't talking like a computer scientist, and I couldn't figure it out. He was using different words, and so we talked for 20 minutes or so, and at the end of the talk, I said, look, you know, I'm dying of curiosity. You are clearly a very smart guy, and you have clearly read my paper in a great deal of detail, but you don't sound like a computer scientist. Who are you? He said, I'm a biologist. So thank you very much. Any questions? I want to know what B stands for in B3. Everybody does. So, you just have no idea what a lunchtime conversation can turn into. So there we were. Merida and I had lunch. We had to give a thing a name, and we were, so B, we were thinking B is, you know, we were working for Boeing at the time, but we couldn't use the name without talking to the lawyers. So there's a B. It has to do with balance. There's another B. Rudy was the senior author. Rudy was several years older than I am. It had been, you know, I had many more publications than I did. So there's another B. And so at the lunchtime, we never did resolve whether there was one of those that made more sense than the rest. So what he likes to say is, the more you think about what the B in B-trees means, the better you understand B-trees. Fair enough. Any other questions? Let's thank the speaker.", "segments": [{"id": 0, "seek": 0, "start": 0.0, "end": 25.12, "text": " So, 40 years ago, that's a long time as Peter said, 40 years ago I remember going to family", "tokens": [50364, 407, 11, 3356, 924, 2057, 11, 300, 311, 257, 938, 565, 382, 6508, 848, 11, 3356, 924, 2057, 286, 1604, 516, 281, 1605, 51620], "temperature": 0.0, "avg_logprob": -0.3402282926771376, "compression_ratio": 1.123456790123457, "no_speech_prob": 0.2876659631729126}, {"id": 1, "seek": 2512, "start": 25.12, "end": 33.6, "text": " reunions and with my brother, who's a medical doctor, we had this old aunt, Aunt Ida.", "tokens": [50364, 14480, 626, 293, 365, 452, 3708, 11, 567, 311, 257, 4625, 4631, 11, 321, 632, 341, 1331, 15654, 11, 17535, 286, 2675, 13, 50788], "temperature": 0.0, "avg_logprob": -0.3574370508608611, "compression_ratio": 1.6629834254143647, "no_speech_prob": 0.8562451601028442}, {"id": 2, "seek": 2512, "start": 33.6, "end": 38.88, "text": " And Aunt Ida would come up to my brother and say, how are things at work, Will?", "tokens": [50788, 400, 17535, 286, 2675, 576, 808, 493, 281, 452, 3708, 293, 584, 11, 577, 366, 721, 412, 589, 11, 3099, 30, 51052], "temperature": 0.0, "avg_logprob": -0.3574370508608611, "compression_ratio": 1.6629834254143647, "no_speech_prob": 0.8562451601028442}, {"id": 3, "seek": 2512, "start": 38.88, "end": 41.88, "text": " And he would say, oh, they're fine Aunt Ida.", "tokens": [51052, 400, 415, 576, 584, 11, 1954, 11, 436, 434, 2489, 17535, 286, 2675, 13, 51202], "temperature": 0.0, "avg_logprob": -0.3574370508608611, "compression_ratio": 1.6629834254143647, "no_speech_prob": 0.8562451601028442}, {"id": 4, "seek": 2512, "start": 41.88, "end": 48.36, "text": " And she would say, you know, I have this pain in my elbow and I have this crick in my back", "tokens": [51202, 400, 750, 576, 584, 11, 291, 458, 11, 286, 362, 341, 1822, 294, 452, 18507, 293, 286, 362, 341, 941, 618, 294, 452, 646, 51526], "temperature": 0.0, "avg_logprob": -0.3574370508608611, "compression_ratio": 1.6629834254143647, "no_speech_prob": 0.8562451601028442}, {"id": 5, "seek": 4836, "start": 48.36, "end": 55.36, "text": " and he would say, well, you know, we'll talk about it tomorrow.", "tokens": [50364, 293, 415, 576, 584, 11, 731, 11, 291, 458, 11, 321, 603, 751, 466, 309, 4153, 13, 50714], "temperature": 0.0, "avg_logprob": -0.27134145848891317, "compression_ratio": 1.5978260869565217, "no_speech_prob": 0.29926368594169617}, {"id": 6, "seek": 4836, "start": 55.36, "end": 60.36, "text": " Aunt Ida would come to me and she would say, how are things at work, Ed?", "tokens": [50714, 17535, 286, 2675, 576, 808, 281, 385, 293, 750, 576, 584, 11, 577, 366, 721, 412, 589, 11, 3977, 30, 50964], "temperature": 0.0, "avg_logprob": -0.27134145848891317, "compression_ratio": 1.5978260869565217, "no_speech_prob": 0.29926368594169617}, {"id": 7, "seek": 4836, "start": 60.36, "end": 64.36, "text": " And I would say, fine, that was the end of the conversation.", "tokens": [50964, 400, 286, 576, 584, 11, 2489, 11, 300, 390, 264, 917, 295, 264, 3761, 13, 51164], "temperature": 0.0, "avg_logprob": -0.27134145848891317, "compression_ratio": 1.5978260869565217, "no_speech_prob": 0.29926368594169617}, {"id": 8, "seek": 4836, "start": 64.36, "end": 70.36, "text": " And I thought to myself at the time, you know, my family is never going to understand what I do.", "tokens": [51164, 400, 286, 1194, 281, 2059, 412, 264, 565, 11, 291, 458, 11, 452, 1605, 307, 1128, 516, 281, 1223, 437, 286, 360, 13, 51464], "temperature": 0.0, "avg_logprob": -0.27134145848891317, "compression_ratio": 1.5978260869565217, "no_speech_prob": 0.29926368594169617}, {"id": 9, "seek": 7036, "start": 70.36, "end": 78.36, "text": " It's such a small group of people and the computers are so expensive and it's such a...", "tokens": [50364, 467, 311, 1270, 257, 1359, 1594, 295, 561, 293, 264, 10807, 366, 370, 5124, 293, 309, 311, 1270, 257, 485, 50764], "temperature": 0.0, "avg_logprob": -0.2148371746665553, "compression_ratio": 1.5632911392405062, "no_speech_prob": 0.34910640120506287}, {"id": 10, "seek": 7036, "start": 78.36, "end": 81.36, "text": " It's fun, I love it, but I'm never going to be.", "tokens": [50764, 467, 311, 1019, 11, 286, 959, 309, 11, 457, 286, 478, 1128, 516, 281, 312, 13, 50914], "temperature": 0.0, "avg_logprob": -0.2148371746665553, "compression_ratio": 1.5632911392405062, "no_speech_prob": 0.34910640120506287}, {"id": 11, "seek": 7036, "start": 81.36, "end": 83.36, "text": " And no one is going to...", "tokens": [50914, 400, 572, 472, 307, 516, 281, 485, 51014], "temperature": 0.0, "avg_logprob": -0.2148371746665553, "compression_ratio": 1.5632911392405062, "no_speech_prob": 0.34910640120506287}, {"id": 12, "seek": 7036, "start": 92.36, "end": 98.36, "text": " So, no one's, I'm never going to get the people in my family to understand what I do.", "tokens": [51464, 407, 11, 572, 472, 311, 11, 286, 478, 1128, 516, 281, 483, 264, 561, 294, 452, 1605, 281, 1223, 437, 286, 360, 13, 51764], "temperature": 0.0, "avg_logprob": -0.2148371746665553, "compression_ratio": 1.5632911392405062, "no_speech_prob": 0.34910640120506287}, {"id": 13, "seek": 9836, "start": 98.36, "end": 106.36, "text": " Now, when my brother and I go to family gatherings, it's, you know, oh, I've got this pain in my elbow, I've got pain in my back,", "tokens": [50364, 823, 11, 562, 452, 3708, 293, 286, 352, 281, 1605, 36247, 11, 309, 311, 11, 291, 458, 11, 1954, 11, 286, 600, 658, 341, 1822, 294, 452, 18507, 11, 286, 600, 658, 1822, 294, 452, 646, 11, 50764], "temperature": 0.0, "avg_logprob": -0.14936683291480654, "compression_ratio": 1.6589147286821706, "no_speech_prob": 0.011363977566361427}, {"id": 14, "seek": 9836, "start": 106.36, "end": 109.36, "text": " my computer won't work, I think I've got a virus.", "tokens": [50764, 452, 3820, 1582, 380, 589, 11, 286, 519, 286, 600, 658, 257, 5752, 13, 50914], "temperature": 0.0, "avg_logprob": -0.14936683291480654, "compression_ratio": 1.6589147286821706, "no_speech_prob": 0.011363977566361427}, {"id": 15, "seek": 9836, "start": 109.36, "end": 116.36, "text": " I'm constantly amazed that now I can talk to people about what I did.", "tokens": [50914, 286, 478, 6460, 20507, 300, 586, 286, 393, 751, 281, 561, 466, 437, 286, 630, 13, 51264], "temperature": 0.0, "avg_logprob": -0.14936683291480654, "compression_ratio": 1.6589147286821706, "no_speech_prob": 0.011363977566361427}, {"id": 16, "seek": 9836, "start": 116.36, "end": 121.36, "text": " I don't do it anymore, as you can see here, perhaps I'm one of the few people who's unaffiliated.", "tokens": [51264, 286, 500, 380, 360, 309, 3602, 11, 382, 291, 393, 536, 510, 11, 4317, 286, 478, 472, 295, 264, 1326, 561, 567, 311, 2002, 602, 2312, 770, 13, 51514], "temperature": 0.0, "avg_logprob": -0.14936683291480654, "compression_ratio": 1.6589147286821706, "no_speech_prob": 0.011363977566361427}, {"id": 17, "seek": 9836, "start": 121.36, "end": 126.36, "text": " And what that means is I wake up every morning and I get to do whatever is fine.", "tokens": [51514, 400, 437, 300, 1355, 307, 286, 6634, 493, 633, 2446, 293, 286, 483, 281, 360, 2035, 307, 2489, 13, 51764], "temperature": 0.0, "avg_logprob": -0.14936683291480654, "compression_ratio": 1.6589147286821706, "no_speech_prob": 0.011363977566361427}, {"id": 18, "seek": 12636, "start": 127.36, "end": 135.36, "text": " So, I'm going to talk about what happened 40 years ago, which was...", "tokens": [50414, 407, 11, 286, 478, 516, 281, 751, 466, 437, 2011, 3356, 924, 2057, 11, 597, 390, 485, 50814], "temperature": 0.0, "avg_logprob": -0.11040079752604166, "compression_ratio": 1.5023696682464456, "no_speech_prob": 0.009325255639851093}, {"id": 19, "seek": 12636, "start": 135.36, "end": 145.36, "text": " First off, I noticed this morning that all the people I might think of mentioning have already been mentioned this morning.", "tokens": [50814, 2386, 766, 11, 286, 5694, 341, 2446, 300, 439, 264, 561, 286, 1062, 519, 295, 18315, 362, 1217, 668, 2835, 341, 2446, 13, 51314], "temperature": 0.0, "avg_logprob": -0.11040079752604166, "compression_ratio": 1.5023696682464456, "no_speech_prob": 0.009325255639851093}, {"id": 20, "seek": 12636, "start": 145.36, "end": 154.36, "text": " So, for example, you might think that it was a pretty small crowd because when I was doing my PhD thesis under Albert Meyer,", "tokens": [51314, 407, 11, 337, 1365, 11, 291, 1062, 519, 300, 309, 390, 257, 1238, 1359, 6919, 570, 562, 286, 390, 884, 452, 14476, 22288, 833, 20812, 47207, 11, 51764], "temperature": 0.0, "avg_logprob": -0.11040079752604166, "compression_ratio": 1.5023696682464456, "no_speech_prob": 0.009325255639851093}, {"id": 21, "seek": 15436, "start": 154.36, "end": 160.36, "text": " as I mentioned, but I was working on Bloom complexity theory, so Manuel Bloom.", "tokens": [50364, 382, 286, 2835, 11, 457, 286, 390, 1364, 322, 25927, 14024, 5261, 11, 370, 34362, 25927, 13, 50664], "temperature": 0.0, "avg_logprob": -0.186063479351741, "compression_ratio": 1.7777777777777777, "no_speech_prob": 0.011059884913265705}, {"id": 22, "seek": 15436, "start": 160.36, "end": 174.36, "text": " And this, I chose a theoretical thesis because I noticed that the guys doing the practical thesis took about twice as long as the guys doing the theoretical thesis.", "tokens": [50664, 400, 341, 11, 286, 5111, 257, 20864, 22288, 570, 286, 5694, 300, 264, 1074, 884, 264, 8496, 22288, 1890, 466, 6091, 382, 938, 382, 264, 1074, 884, 264, 20864, 22288, 13, 51364], "temperature": 0.0, "avg_logprob": -0.186063479351741, "compression_ratio": 1.7777777777777777, "no_speech_prob": 0.011059884913265705}, {"id": 23, "seek": 15436, "start": 174.36, "end": 177.36, "text": " And I was really inquiring it out of graduate school.", "tokens": [51364, 400, 286, 390, 534, 13570, 5057, 309, 484, 295, 8080, 1395, 13, 51514], "temperature": 0.0, "avg_logprob": -0.186063479351741, "compression_ratio": 1.7777777777777777, "no_speech_prob": 0.011059884913265705}, {"id": 24, "seek": 15436, "start": 177.36, "end": 183.36, "text": " So, I decided to do a theoretical thesis, it was only 50 pages, you know, it was nice.", "tokens": [51514, 407, 11, 286, 3047, 281, 360, 257, 20864, 22288, 11, 309, 390, 787, 2625, 7183, 11, 291, 458, 11, 309, 390, 1481, 13, 51814], "temperature": 0.0, "avg_logprob": -0.186063479351741, "compression_ratio": 1.7777777777777777, "no_speech_prob": 0.011059884913265705}, {"id": 25, "seek": 18336, "start": 183.36, "end": 192.36, "text": " But the theorems were of the form for all models of computation, for all majors of complexity, blah, blah, blah.", "tokens": [50364, 583, 264, 10299, 2592, 645, 295, 264, 1254, 337, 439, 5245, 295, 24903, 11, 337, 439, 31770, 295, 14024, 11, 12288, 11, 12288, 11, 12288, 13, 50814], "temperature": 0.0, "avg_logprob": -0.15378846178998004, "compression_ratio": 1.6901408450704225, "no_speech_prob": 0.006409692578017712}, {"id": 26, "seek": 18336, "start": 192.36, "end": 199.36, "text": " So, for example, there's an arbitrarily, you can write an arbitrarily bad program to do anything.", "tokens": [50814, 407, 11, 337, 1365, 11, 456, 311, 364, 19071, 3289, 11, 291, 393, 2464, 364, 19071, 3289, 1578, 1461, 281, 360, 1340, 13, 51164], "temperature": 0.0, "avg_logprob": -0.15378846178998004, "compression_ratio": 1.6901408450704225, "no_speech_prob": 0.006409692578017712}, {"id": 27, "seek": 18336, "start": 199.36, "end": 203.36, "text": " Okay, programmers intuition, absolutely true.", "tokens": [51164, 1033, 11, 41504, 24002, 11, 3122, 2074, 13, 51364], "temperature": 0.0, "avg_logprob": -0.15378846178998004, "compression_ratio": 1.6901408450704225, "no_speech_prob": 0.006409692578017712}, {"id": 28, "seek": 18336, "start": 203.36, "end": 209.36, "text": " Or there are some functions that are just plain hard to compute, no matter how you do it, they're hard.", "tokens": [51364, 1610, 456, 366, 512, 6828, 300, 366, 445, 11121, 1152, 281, 14722, 11, 572, 1871, 577, 291, 360, 309, 11, 436, 434, 1152, 13, 51664], "temperature": 0.0, "avg_logprob": -0.15378846178998004, "compression_ratio": 1.6901408450704225, "no_speech_prob": 0.006409692578017712}, {"id": 29, "seek": 20936, "start": 209.36, "end": 219.36, "text": " But even after finishing my thesis in that theory, as far as I know, the only really interesting theorem in that theory was proved by Manuel Bloom himself,", "tokens": [50364, 583, 754, 934, 12693, 452, 22288, 294, 300, 5261, 11, 382, 1400, 382, 286, 458, 11, 264, 787, 534, 1880, 20904, 294, 300, 5261, 390, 14617, 538, 34362, 25927, 3647, 11, 50864], "temperature": 0.0, "avg_logprob": -0.0910413553426554, "compression_ratio": 1.705128205128205, "no_speech_prob": 0.013470265083014965}, {"id": 30, "seek": 20936, "start": 219.36, "end": 227.36, "text": " which was that you have to be very careful with the notion best program because there are some functions that don't have one.", "tokens": [50864, 597, 390, 300, 291, 362, 281, 312, 588, 5026, 365, 264, 10710, 1151, 1461, 570, 456, 366, 512, 6828, 300, 500, 380, 362, 472, 13, 51264], "temperature": 0.0, "avg_logprob": -0.0910413553426554, "compression_ratio": 1.705128205128205, "no_speech_prob": 0.013470265083014965}, {"id": 31, "seek": 20936, "start": 227.36, "end": 236.36, "text": " In the sense that if I exhibit a program and say, this is the best program, then someone else can come along and say,", "tokens": [51264, 682, 264, 2020, 300, 498, 286, 20487, 257, 1461, 293, 584, 11, 341, 307, 264, 1151, 1461, 11, 550, 1580, 1646, 393, 808, 2051, 293, 584, 11, 51714], "temperature": 0.0, "avg_logprob": -0.0910413553426554, "compression_ratio": 1.705128205128205, "no_speech_prob": 0.013470265083014965}, {"id": 32, "seek": 23636, "start": 236.36, "end": 239.36, "text": " here's a proof that there is a better program.", "tokens": [50364, 510, 311, 257, 8177, 300, 456, 307, 257, 1101, 1461, 13, 50514], "temperature": 0.0, "avg_logprob": -0.12734867334365846, "compression_ratio": 1.4816753926701571, "no_speech_prob": 0.013918556272983551}, {"id": 33, "seek": 23636, "start": 239.36, "end": 243.36, "text": " I can't show you what it is, but here's the proof.", "tokens": [50514, 286, 393, 380, 855, 291, 437, 309, 307, 11, 457, 510, 311, 264, 8177, 13, 50714], "temperature": 0.0, "avg_logprob": -0.12734867334365846, "compression_ratio": 1.4816753926701571, "no_speech_prob": 0.013918556272983551}, {"id": 34, "seek": 23636, "start": 243.36, "end": 249.36, "text": " So, but I was a radio amateur as a kid, I loved electronics.", "tokens": [50714, 407, 11, 457, 286, 390, 257, 6477, 29339, 382, 257, 1636, 11, 286, 4333, 20611, 13, 51014], "temperature": 0.0, "avg_logprob": -0.12734867334365846, "compression_ratio": 1.4816753926701571, "no_speech_prob": 0.013918556272983551}, {"id": 35, "seek": 23636, "start": 249.36, "end": 262.36, "text": " I was a very practical guy, and living in a world where Ackerman's function and primitive recursion kind of melded together,", "tokens": [51014, 286, 390, 257, 588, 8496, 2146, 11, 293, 2647, 294, 257, 1002, 689, 316, 547, 11821, 311, 2445, 293, 28540, 20560, 313, 733, 295, 4795, 9207, 1214, 11, 51664], "temperature": 0.0, "avg_logprob": -0.12734867334365846, "compression_ratio": 1.4816753926701571, "no_speech_prob": 0.013918556272983551}, {"id": 36, "seek": 26236, "start": 262.36, "end": 266.36, "text": " just wasn't a place where I wanted to be really.", "tokens": [50364, 445, 2067, 380, 257, 1081, 689, 286, 1415, 281, 312, 534, 13, 50564], "temperature": 0.0, "avg_logprob": -0.11522924447361427, "compression_ratio": 1.5853658536585367, "no_speech_prob": 0.01561563927680254}, {"id": 37, "seek": 26236, "start": 266.36, "end": 277.36, "text": " So when I finished my PhD, I went to Boeing, the aircraft company, and they had a research lab there, a wonderful place where people were analyzing the theory of the moon,", "tokens": [50564, 407, 562, 286, 4335, 452, 14476, 11, 286, 1437, 281, 30831, 11, 264, 9465, 2237, 11, 293, 436, 632, 257, 2132, 2715, 456, 11, 257, 3715, 1081, 689, 561, 645, 23663, 264, 5261, 295, 264, 7135, 11, 51114], "temperature": 0.0, "avg_logprob": -0.11522924447361427, "compression_ratio": 1.5853658536585367, "no_speech_prob": 0.01561563927680254}, {"id": 38, "seek": 26236, "start": 277.36, "end": 286.36, "text": " so how the moon orbits the earth, and they were studying aerodynamics, and it was all fascinating stuff.", "tokens": [51114, 370, 577, 264, 7135, 43522, 264, 4120, 11, 293, 436, 645, 7601, 11207, 35483, 11, 293, 309, 390, 439, 10343, 1507, 13, 51564], "temperature": 0.0, "avg_logprob": -0.11522924447361427, "compression_ratio": 1.5853658536585367, "no_speech_prob": 0.01561563927680254}, {"id": 39, "seek": 28636, "start": 286.36, "end": 297.36, "text": " So that's where I met Rudy Bayer, and we did B-trees together at Boeing, but at that point Boeing was running out of money,", "tokens": [50364, 407, 300, 311, 689, 286, 1131, 38690, 7840, 260, 11, 293, 321, 630, 363, 12, 3599, 279, 1214, 412, 30831, 11, 457, 412, 300, 935, 30831, 390, 2614, 484, 295, 1460, 11, 50914], "temperature": 0.0, "avg_logprob": -0.1396434016344024, "compression_ratio": 1.5784313725490196, "no_speech_prob": 0.7789696455001831}, {"id": 40, "seek": 28636, "start": 297.36, "end": 306.36, "text": " and so I could see the handwriting on the wall that our lab was going to disappear shortly.", "tokens": [50914, 293, 370, 286, 727, 536, 264, 39179, 322, 264, 2929, 300, 527, 2715, 390, 516, 281, 11596, 13392, 13, 51364], "temperature": 0.0, "avg_logprob": -0.1396434016344024, "compression_ratio": 1.5784313725490196, "no_speech_prob": 0.7789696455001831}, {"id": 41, "seek": 28636, "start": 306.36, "end": 315.36, "text": " They were reducing, I think maybe three employees out of four were fired over a one-year period at Boeing.", "tokens": [51364, 814, 645, 12245, 11, 286, 519, 1310, 1045, 6619, 484, 295, 1451, 645, 11777, 670, 257, 472, 12, 5294, 2896, 412, 30831, 13, 51814], "temperature": 0.0, "avg_logprob": -0.1396434016344024, "compression_ratio": 1.5784313725490196, "no_speech_prob": 0.7789696455001831}, {"id": 42, "seek": 31536, "start": 315.36, "end": 328.36, "text": " Rudy and I were both on a job hunt, and one place, I had a paper in my pocket, the B-tree paper, that had been rejected for publication,", "tokens": [50364, 38690, 293, 286, 645, 1293, 322, 257, 1691, 12454, 11, 293, 472, 1081, 11, 286, 632, 257, 3035, 294, 452, 8963, 11, 264, 363, 12, 83, 701, 3035, 11, 300, 632, 668, 15749, 337, 19953, 11, 51014], "temperature": 0.0, "avg_logprob": -0.1419101597107563, "compression_ratio": 1.5924170616113744, "no_speech_prob": 0.030799292027950287}, {"id": 43, "seek": 31536, "start": 328.36, "end": 335.36, "text": " but you know, it was kind of interesting stuff, and so when people asked me to talk at, say, a university like Stanford,", "tokens": [51014, 457, 291, 458, 11, 309, 390, 733, 295, 1880, 1507, 11, 293, 370, 562, 561, 2351, 385, 281, 751, 412, 11, 584, 11, 257, 5454, 411, 20374, 11, 51364], "temperature": 0.0, "avg_logprob": -0.1419101597107563, "compression_ratio": 1.5924170616113744, "no_speech_prob": 0.030799292027950287}, {"id": 44, "seek": 31536, "start": 335.36, "end": 342.36, "text": " you know, I pulled out the B-tree paper and I stood up and gave a B-tree talk,", "tokens": [51364, 291, 458, 11, 286, 7373, 484, 264, 363, 12, 83, 701, 3035, 293, 286, 9371, 493, 293, 2729, 257, 363, 12, 83, 701, 751, 11, 51714], "temperature": 0.0, "avg_logprob": -0.1419101597107563, "compression_ratio": 1.5924170616113744, "no_speech_prob": 0.030799292027950287}, {"id": 45, "seek": 34236, "start": 342.36, "end": 349.36, "text": " and as it happened at Stanford, Don Teuth was in the audience, and he was at that moment writing volume three, sorting and searching,", "tokens": [50364, 293, 382, 309, 2011, 412, 20374, 11, 1468, 1989, 2910, 390, 294, 264, 4034, 11, 293, 415, 390, 412, 300, 1623, 3579, 5523, 1045, 11, 32411, 293, 10808, 11, 50714], "temperature": 0.0, "avg_logprob": -0.18289216010125128, "compression_ratio": 1.6129032258064515, "no_speech_prob": 0.032313548028469086}, {"id": 46, "seek": 34236, "start": 349.36, "end": 358.36, "text": " and so he said, he came up after Rudy, and he said, that was kind of interesting, you know, I'm writing this book,", "tokens": [50714, 293, 370, 415, 848, 11, 415, 1361, 493, 934, 38690, 11, 293, 415, 848, 11, 300, 390, 733, 295, 1880, 11, 291, 458, 11, 286, 478, 3579, 341, 1446, 11, 51164], "temperature": 0.0, "avg_logprob": -0.18289216010125128, "compression_ratio": 1.6129032258064515, "no_speech_prob": 0.032313548028469086}, {"id": 47, "seek": 34236, "start": 358.36, "end": 365.36, "text": " can you give me a reference to your paper, and I said, you know, we've been rejected for publication,", "tokens": [51164, 393, 291, 976, 385, 257, 6408, 281, 428, 3035, 11, 293, 286, 848, 11, 291, 458, 11, 321, 600, 668, 15749, 337, 19953, 11, 51514], "temperature": 0.0, "avg_logprob": -0.18289216010125128, "compression_ratio": 1.6129032258064515, "no_speech_prob": 0.032313548028469086}, {"id": 48, "seek": 36536, "start": 365.36, "end": 372.36, "text": " because in the communications of the ACM, the computing techniques editor is a hashing fan,", "tokens": [50364, 570, 294, 264, 15163, 295, 264, 8157, 44, 11, 264, 15866, 7512, 9839, 307, 257, 575, 571, 3429, 11, 50714], "temperature": 0.0, "avg_logprob": -0.13449404456398703, "compression_ratio": 1.5, "no_speech_prob": 0.4524405300617218}, {"id": 49, "seek": 36536, "start": 372.36, "end": 380.36, "text": " and he thought that there were too many indexing papers recently, and so, you know, you're out of luck.", "tokens": [50714, 293, 415, 1194, 300, 456, 645, 886, 867, 8186, 278, 10577, 3938, 11, 293, 370, 11, 291, 458, 11, 291, 434, 484, 295, 3668, 13, 51114], "temperature": 0.0, "avg_logprob": -0.13449404456398703, "compression_ratio": 1.5, "no_speech_prob": 0.4524405300617218}, {"id": 50, "seek": 36536, "start": 380.36, "end": 388.36, "text": " So Don said, oh, I've got this journal in my back pocket, just give me the paper, it will publish it, so I can point my book at it,", "tokens": [51114, 407, 1468, 848, 11, 1954, 11, 286, 600, 658, 341, 6708, 294, 452, 646, 8963, 11, 445, 976, 385, 264, 3035, 11, 309, 486, 11374, 309, 11, 370, 286, 393, 935, 452, 1446, 412, 309, 11, 51514], "temperature": 0.0, "avg_logprob": -0.13449404456398703, "compression_ratio": 1.5, "no_speech_prob": 0.4524405300617218}, {"id": 51, "seek": 38836, "start": 388.36, "end": 395.36, "text": " and by the way, give me all your simulations and everything, so all the card decks, all the printed listings went to Don,", "tokens": [50364, 293, 538, 264, 636, 11, 976, 385, 439, 428, 35138, 293, 1203, 11, 370, 439, 264, 2920, 32607, 11, 439, 264, 13567, 45615, 1437, 281, 1468, 11, 50714], "temperature": 0.0, "avg_logprob": -0.11958861800859559, "compression_ratio": 1.6693877551020408, "no_speech_prob": 0.3376688063144684}, {"id": 52, "seek": 38836, "start": 395.36, "end": 401.36, "text": " and three months later, they came back, and then, of course, when volume three came out, I read volume three,", "tokens": [50714, 293, 1045, 2493, 1780, 11, 436, 1361, 646, 11, 293, 550, 11, 295, 1164, 11, 562, 5523, 1045, 1361, 484, 11, 286, 1401, 5523, 1045, 11, 51014], "temperature": 0.0, "avg_logprob": -0.11958861800859559, "compression_ratio": 1.6693877551020408, "no_speech_prob": 0.3376688063144684}, {"id": 53, "seek": 38836, "start": 401.36, "end": 407.36, "text": " and I said, oh, that's something we should have written, if only we could write that well.", "tokens": [51014, 293, 286, 848, 11, 1954, 11, 300, 311, 746, 321, 820, 362, 3720, 11, 498, 787, 321, 727, 2464, 300, 731, 13, 51314], "temperature": 0.0, "avg_logprob": -0.11958861800859559, "compression_ratio": 1.6693877551020408, "no_speech_prob": 0.3376688063144684}, {"id": 54, "seek": 38836, "start": 407.36, "end": 413.36, "text": " So anyway, I was on this job search, and one of the places I went was Yale University,", "tokens": [51314, 407, 4033, 11, 286, 390, 322, 341, 1691, 3164, 11, 293, 472, 295, 264, 3190, 286, 1437, 390, 26711, 3535, 11, 51614], "temperature": 0.0, "avg_logprob": -0.11958861800859559, "compression_ratio": 1.6693877551020408, "no_speech_prob": 0.3376688063144684}, {"id": 55, "seek": 41336, "start": 413.36, "end": 423.36, "text": " where there was Peter recruiting like crazy with Ned Irons, and, you know, they seemed to think I was interesting,", "tokens": [50364, 689, 456, 390, 6508, 25987, 411, 3219, 365, 31355, 9151, 892, 11, 293, 11, 291, 458, 11, 436, 6576, 281, 519, 286, 390, 1880, 11, 50864], "temperature": 0.0, "avg_logprob": -0.13486630207783468, "compression_ratio": 1.5175879396984924, "no_speech_prob": 0.35928741097450256}, {"id": 56, "seek": 41336, "start": 423.36, "end": 430.36, "text": " and they offered me an assistant professor position, and it sounded really terrific until I asked,", "tokens": [50864, 293, 436, 8059, 385, 364, 10994, 8304, 2535, 11, 293, 309, 17714, 534, 20899, 1826, 286, 2351, 11, 51214], "temperature": 0.0, "avg_logprob": -0.13486630207783468, "compression_ratio": 1.5175879396984924, "no_speech_prob": 0.35928741097450256}, {"id": 57, "seek": 41336, "start": 430.36, "end": 436.36, "text": " and by the way, it's kind of hard to park around here, do I get a reserve parking space?", "tokens": [51214, 293, 538, 264, 636, 11, 309, 311, 733, 295, 1152, 281, 3884, 926, 510, 11, 360, 286, 483, 257, 17824, 9893, 1901, 30, 51514], "temperature": 0.0, "avg_logprob": -0.13486630207783468, "compression_ratio": 1.5175879396984924, "no_speech_prob": 0.35928741097450256}, {"id": 58, "seek": 43636, "start": 436.36, "end": 443.36, "text": " Oh, and then I understood, you know, the importance of my position, really.", "tokens": [50364, 876, 11, 293, 550, 286, 7320, 11, 291, 458, 11, 264, 7379, 295, 452, 2535, 11, 534, 13, 50714], "temperature": 0.0, "avg_logprob": -0.24138699849446615, "compression_ratio": 1.451086956521739, "no_speech_prob": 0.05034927278757095}, {"id": 59, "seek": 43636, "start": 443.36, "end": 449.36, "text": " That's the most valuable resource in the universe?", "tokens": [50714, 663, 311, 264, 881, 8263, 7684, 294, 264, 6445, 30, 51014], "temperature": 0.0, "avg_logprob": -0.24138699849446615, "compression_ratio": 1.451086956521739, "no_speech_prob": 0.05034927278757095}, {"id": 60, "seek": 43636, "start": 449.36, "end": 453.36, "text": " Exactly, well, I can't understand that.", "tokens": [51014, 7587, 11, 731, 11, 286, 393, 380, 1223, 300, 13, 51214], "temperature": 0.0, "avg_logprob": -0.24138699849446615, "compression_ratio": 1.451086956521739, "no_speech_prob": 0.05034927278757095}, {"id": 61, "seek": 43636, "start": 453.36, "end": 462.36, "text": " So, but in the end, that job search led me to the Iraq's park, where a lot of my friends were going,", "tokens": [51214, 407, 11, 457, 294, 264, 917, 11, 300, 1691, 3164, 4684, 385, 281, 264, 11818, 311, 3884, 11, 689, 257, 688, 295, 452, 1855, 645, 516, 11, 51664], "temperature": 0.0, "avg_logprob": -0.24138699849446615, "compression_ratio": 1.451086956521739, "no_speech_prob": 0.05034927278757095}, {"id": 62, "seek": 46236, "start": 463.36, "end": 469.36, "text": " it was parking, yes, they had parking in park, that's right.", "tokens": [50414, 309, 390, 9893, 11, 2086, 11, 436, 632, 9893, 294, 3884, 11, 300, 311, 558, 13, 50714], "temperature": 0.0, "avg_logprob": -0.17481072369743797, "compression_ratio": 1.6486486486486487, "no_speech_prob": 0.08377940952777863}, {"id": 63, "seek": 46236, "start": 469.36, "end": 476.36, "text": " There was lots of it, and you were 24 hours a day, there were as many people there at one o'clock in the morning", "tokens": [50714, 821, 390, 3195, 295, 309, 11, 293, 291, 645, 4022, 2496, 257, 786, 11, 456, 645, 382, 867, 561, 456, 412, 472, 277, 6, 9023, 294, 264, 2446, 51064], "temperature": 0.0, "avg_logprob": -0.17481072369743797, "compression_ratio": 1.6486486486486487, "no_speech_prob": 0.08377940952777863}, {"id": 64, "seek": 46236, "start": 476.36, "end": 479.36, "text": " as there were at eight o'clock in the morning.", "tokens": [51064, 382, 456, 645, 412, 3180, 277, 6, 9023, 294, 264, 2446, 13, 51214], "temperature": 0.0, "avg_logprob": -0.17481072369743797, "compression_ratio": 1.6486486486486487, "no_speech_prob": 0.08377940952777863}, {"id": 65, "seek": 46236, "start": 479.36, "end": 485.36, "text": " It was a really interesting place, but Peter and I had stayed in touch a little bit,", "tokens": [51214, 467, 390, 257, 534, 1880, 1081, 11, 457, 6508, 293, 286, 632, 9181, 294, 2557, 257, 707, 857, 11, 51514], "temperature": 0.0, "avg_logprob": -0.17481072369743797, "compression_ratio": 1.6486486486486487, "no_speech_prob": 0.08377940952777863}, {"id": 66, "seek": 48536, "start": 485.36, "end": 492.36, "text": " and just before I was scheduled to go on a family camping vacation in 1973,", "tokens": [50364, 293, 445, 949, 286, 390, 15678, 281, 352, 322, 257, 1605, 19470, 12830, 294, 33530, 11, 50714], "temperature": 0.0, "avg_logprob": -0.09447775681813558, "compression_ratio": 1.3701657458563536, "no_speech_prob": 0.025128647685050964}, {"id": 67, "seek": 48536, "start": 492.36, "end": 498.36, "text": " I got this note from Peter, said, considering the following problem,", "tokens": [50714, 286, 658, 341, 3637, 490, 6508, 11, 848, 11, 8079, 264, 3480, 1154, 11, 51014], "temperature": 0.0, "avg_logprob": -0.09447775681813558, "compression_ratio": 1.3701657458563536, "no_speech_prob": 0.025128647685050964}, {"id": 68, "seek": 48536, "start": 498.36, "end": 509.36, "text": " suppose we send our astronauts off to some distant moon carrying a copy of the Encyclopedia Britannica,", "tokens": [51014, 7297, 321, 2845, 527, 28273, 766, 281, 512, 17275, 7135, 9792, 257, 5055, 295, 264, 2193, 34080, 47795, 4760, 969, 2262, 11, 51564], "temperature": 0.0, "avg_logprob": -0.09447775681813558, "compression_ratio": 1.3701657458563536, "no_speech_prob": 0.025128647685050964}, {"id": 69, "seek": 50936, "start": 509.36, "end": 517.36, "text": " and years later, and there's this really slow, crappy radio link to these guys,", "tokens": [50364, 293, 924, 1780, 11, 293, 456, 311, 341, 534, 2964, 11, 36531, 6477, 2113, 281, 613, 1074, 11, 50764], "temperature": 0.0, "avg_logprob": -0.08889294633961688, "compression_ratio": 1.83, "no_speech_prob": 0.01404337678104639}, {"id": 70, "seek": 50936, "start": 517.36, "end": 522.36, "text": " and years later, Encyclopedia Britannica comes out with a new version,", "tokens": [50764, 293, 924, 1780, 11, 2193, 34080, 47795, 4760, 969, 2262, 1487, 484, 365, 257, 777, 3037, 11, 51014], "temperature": 0.0, "avg_logprob": -0.08889294633961688, "compression_ratio": 1.83, "no_speech_prob": 0.01404337678104639}, {"id": 71, "seek": 50936, "start": 522.36, "end": 529.36, "text": " and so we want to send them the new Encyclopedia, only, you know, we've got this crappy little radio link to do it with.", "tokens": [51014, 293, 370, 321, 528, 281, 2845, 552, 264, 777, 2193, 34080, 47795, 11, 787, 11, 291, 458, 11, 321, 600, 658, 341, 36531, 707, 6477, 2113, 281, 360, 309, 365, 13, 51364], "temperature": 0.0, "avg_logprob": -0.08889294633961688, "compression_ratio": 1.83, "no_speech_prob": 0.01404337678104639}, {"id": 72, "seek": 50936, "start": 529.36, "end": 537.36, "text": " So, imagine that you want to encode the new Encyclopedia with respect to the old Encyclopedia,", "tokens": [51364, 407, 11, 3811, 300, 291, 528, 281, 2058, 1429, 264, 777, 2193, 34080, 47795, 365, 3104, 281, 264, 1331, 2193, 34080, 47795, 11, 51764], "temperature": 0.0, "avg_logprob": -0.08889294633961688, "compression_ratio": 1.83, "no_speech_prob": 0.01404337678104639}, {"id": 73, "seek": 53736, "start": 537.36, "end": 544.36, "text": " and you can do it with two different kinds of code words, one, a letter,", "tokens": [50364, 293, 291, 393, 360, 309, 365, 732, 819, 3685, 295, 3089, 2283, 11, 472, 11, 257, 5063, 11, 50714], "temperature": 0.0, "avg_logprob": -0.14159635261253076, "compression_ratio": 1.7908163265306123, "no_speech_prob": 0.004164563026279211}, {"id": 74, "seek": 53736, "start": 544.36, "end": 552.36, "text": " or so any one of the, I don't know, 50 or 100 letters, or the other, a substring reference,", "tokens": [50714, 420, 370, 604, 472, 295, 264, 11, 286, 500, 380, 458, 11, 2625, 420, 2319, 7825, 11, 420, 264, 661, 11, 257, 4594, 2937, 6408, 11, 51114], "temperature": 0.0, "avg_logprob": -0.14159635261253076, "compression_ratio": 1.7908163265306123, "no_speech_prob": 0.004164563026279211}, {"id": 75, "seek": 53736, "start": 552.36, "end": 559.36, "text": " where you say from position so and so to position so and so in the old Encyclopedia,", "tokens": [51114, 689, 291, 584, 490, 2535, 370, 293, 370, 281, 2535, 370, 293, 370, 294, 264, 1331, 2193, 34080, 47795, 11, 51464], "temperature": 0.0, "avg_logprob": -0.14159635261253076, "compression_ratio": 1.7908163265306123, "no_speech_prob": 0.004164563026279211}, {"id": 76, "seek": 53736, "start": 559.36, "end": 566.36, "text": " so, you know, A, B, Z, and then the substring, and then the E, F, and then this substring, and so on.", "tokens": [51464, 370, 11, 291, 458, 11, 316, 11, 363, 11, 1176, 11, 293, 550, 264, 4594, 2937, 11, 293, 550, 264, 462, 11, 479, 11, 293, 550, 341, 4594, 2937, 11, 293, 370, 322, 13, 51814], "temperature": 0.0, "avg_logprob": -0.14159635261253076, "compression_ratio": 1.7908163265306123, "no_speech_prob": 0.004164563026279211}, {"id": 77, "seek": 56636, "start": 566.36, "end": 576.36, "text": " You said, I think I found a way to find the optimum encoding for that problem in linear time,", "tokens": [50364, 509, 848, 11, 286, 519, 286, 1352, 257, 636, 281, 915, 264, 39326, 43430, 337, 300, 1154, 294, 8213, 565, 11, 50864], "temperature": 0.0, "avg_logprob": -0.19221328553699313, "compression_ratio": 1.4904458598726114, "no_speech_prob": 0.002289548981934786}, {"id": 78, "seek": 56636, "start": 576.36, "end": 581.36, "text": " time linear in the length, in the sum of lengths, through Encyclopedias.", "tokens": [50864, 565, 8213, 294, 264, 4641, 11, 294, 264, 2408, 295, 26329, 11, 807, 2193, 34080, 27277, 4609, 13, 51114], "temperature": 0.0, "avg_logprob": -0.19221328553699313, "compression_ratio": 1.4904458598726114, "no_speech_prob": 0.002289548981934786}, {"id": 79, "seek": 56636, "start": 581.36, "end": 593.36, "text": " And by the way, I think my solution to this problem involves a try,", "tokens": [51114, 400, 538, 264, 636, 11, 286, 519, 452, 3827, 281, 341, 1154, 11626, 257, 853, 11, 51714], "temperature": 0.0, "avg_logprob": -0.19221328553699313, "compression_ratio": 1.4904458598726114, "no_speech_prob": 0.002289548981934786}, {"id": 80, "seek": 59336, "start": 594.36, "end": 601.36, "text": " so a tree of suffixes and things.", "tokens": [50414, 370, 257, 4230, 295, 3889, 36005, 293, 721, 13, 50764], "temperature": 0.0, "avg_logprob": -0.14400954305389782, "compression_ratio": 1.6084656084656084, "no_speech_prob": 0.004533798433840275}, {"id": 81, "seek": 59336, "start": 601.36, "end": 609.36, "text": " And so it involves this auxiliary structure that helps me, that I built during the reading of the old Encyclopedia,", "tokens": [50764, 400, 370, 309, 11626, 341, 43741, 3877, 300, 3665, 385, 11, 300, 286, 3094, 1830, 264, 3760, 295, 264, 1331, 2193, 34080, 47795, 11, 51164], "temperature": 0.0, "avg_logprob": -0.14400954305389782, "compression_ratio": 1.6084656084656084, "no_speech_prob": 0.004533798433840275}, {"id": 82, "seek": 59336, "start": 609.36, "end": 613.36, "text": " and then I use it for the encoding of the new Encyclopedia.", "tokens": [51164, 293, 550, 286, 764, 309, 337, 264, 43430, 295, 264, 777, 2193, 34080, 47795, 13, 51364], "temperature": 0.0, "avg_logprob": -0.14400954305389782, "compression_ratio": 1.6084656084656084, "no_speech_prob": 0.004533798433840275}, {"id": 83, "seek": 59336, "start": 613.36, "end": 621.36, "text": " And then the next day I went on vacation, and my first reaction was, hey, that's not possible.", "tokens": [51364, 400, 550, 264, 958, 786, 286, 1437, 322, 12830, 11, 293, 452, 700, 5480, 390, 11, 4177, 11, 300, 311, 406, 1944, 13, 51764], "temperature": 0.0, "avg_logprob": -0.14400954305389782, "compression_ratio": 1.6084656084656084, "no_speech_prob": 0.004533798433840275}, {"id": 84, "seek": 62136, "start": 621.36, "end": 627.36, "text": " Come on. It's not even a factor of log ending here.", "tokens": [50364, 2492, 322, 13, 467, 311, 406, 754, 257, 5952, 295, 3565, 8121, 510, 13, 50664], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 85, "seek": 62136, "start": 627.36, "end": 631.36, "text": " How could you possibly do this?", "tokens": [50664, 1012, 727, 291, 6264, 360, 341, 30, 50864], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 86, "seek": 62136, "start": 631.36, "end": 634.36, "text": " And so for two weeks, my family must have wondered where I was.", "tokens": [50864, 400, 370, 337, 732, 3259, 11, 452, 1605, 1633, 362, 17055, 689, 286, 390, 13, 51014], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 87, "seek": 62136, "start": 634.36, "end": 639.36, "text": " I mean, I was there, but I wasn't there, because I was puzzling over this thing.", "tokens": [51014, 286, 914, 11, 286, 390, 456, 11, 457, 286, 2067, 380, 456, 11, 570, 286, 390, 18741, 1688, 670, 341, 551, 13, 51264], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 88, "seek": 62136, "start": 639.36, "end": 645.36, "text": " And I didn't see a paper, I didn't see any of the stuff that Vaughn saw.", "tokens": [51264, 400, 286, 994, 380, 536, 257, 3035, 11, 286, 994, 380, 536, 604, 295, 264, 1507, 300, 16822, 1984, 77, 1866, 13, 51564], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 89, "seek": 62136, "start": 645.36, "end": 647.36, "text": " I just saw the result.", "tokens": [51564, 286, 445, 1866, 264, 1874, 13, 51664], "temperature": 0.0, "avg_logprob": -0.13575156211853026, "compression_ratio": 1.5355450236966826, "no_speech_prob": 0.01366353314369917}, {"id": 90, "seek": 64736, "start": 648.36, "end": 652.36, "text": " But I thought, you know, I'm not that much younger than Peter.", "tokens": [50414, 583, 286, 1194, 11, 291, 458, 11, 286, 478, 406, 300, 709, 7037, 813, 6508, 13, 50614], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 91, "seek": 64736, "start": 652.36, "end": 655.36, "text": " I should be able to work this out.", "tokens": [50614, 286, 820, 312, 1075, 281, 589, 341, 484, 13, 50764], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 92, "seek": 64736, "start": 655.36, "end": 660.36, "text": " I mean, knowing that something is true is a tremendous head start.", "tokens": [50764, 286, 914, 11, 5276, 300, 746, 307, 2074, 307, 257, 10048, 1378, 722, 13, 51014], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 93, "seek": 64736, "start": 660.36, "end": 663.36, "text": " So, you know, come on, you can do this.", "tokens": [51014, 407, 11, 291, 458, 11, 808, 322, 11, 291, 393, 360, 341, 13, 51164], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 94, "seek": 64736, "start": 663.36, "end": 666.36, "text": " And at the end of the two weeks, I figured it out.", "tokens": [51164, 400, 412, 264, 917, 295, 264, 732, 3259, 11, 286, 8932, 309, 484, 13, 51314], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 95, "seek": 64736, "start": 666.36, "end": 668.36, "text": " It was brilliant.", "tokens": [51314, 467, 390, 10248, 13, 51414], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 96, "seek": 64736, "start": 668.36, "end": 670.36, "text": " And so I wrote Peter and know what he said.", "tokens": [51414, 400, 370, 286, 4114, 6508, 293, 458, 437, 415, 848, 13, 51514], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 97, "seek": 64736, "start": 670.36, "end": 674.36, "text": " Peter, you know, I studied and studied and studied over this thing,", "tokens": [51514, 6508, 11, 291, 458, 11, 286, 9454, 293, 9454, 293, 9454, 670, 341, 551, 11, 51714], "temperature": 0.0, "avg_logprob": -0.12861506144205728, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0194107573479414}, {"id": 98, "seek": 67436, "start": 674.36, "end": 677.36, "text": " and I figured it out. You're brilliant.", "tokens": [50364, 293, 286, 8932, 309, 484, 13, 509, 434, 10248, 13, 50514], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 99, "seek": 67436, "start": 677.36, "end": 680.36, "text": " This is what you must have done.", "tokens": [50514, 639, 307, 437, 291, 1633, 362, 1096, 13, 50664], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 100, "seek": 67436, "start": 680.36, "end": 688.36, "text": " Peter wrote back and said, er, no, actually, it's different.", "tokens": [50664, 6508, 4114, 646, 293, 848, 11, 1189, 11, 572, 11, 767, 11, 309, 311, 819, 13, 51064], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 101, "seek": 67436, "start": 688.36, "end": 692.36, "text": " And in some ways, what you wrote to me is better,", "tokens": [51064, 400, 294, 512, 2098, 11, 437, 291, 4114, 281, 385, 307, 1101, 11, 51264], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 102, "seek": 67436, "start": 692.36, "end": 696.36, "text": " because, for example, well, one way for humanities,", "tokens": [51264, 570, 11, 337, 1365, 11, 731, 11, 472, 636, 337, 36140, 11, 51464], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 103, "seek": 67436, "start": 696.36, "end": 701.36, "text": " it's got one fewer pointer per note.", "tokens": [51464, 309, 311, 658, 472, 13366, 23918, 680, 3637, 13, 51714], "temperature": 0.0, "avg_logprob": -0.20448924273979374, "compression_ratio": 1.439153439153439, "no_speech_prob": 0.002387561136856675}, {"id": 104, "seek": 70136, "start": 701.36, "end": 704.36, "text": " And it works from left to right.", "tokens": [50364, 400, 309, 1985, 490, 1411, 281, 558, 13, 50514], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 105, "seek": 70136, "start": 704.36, "end": 708.36, "text": " So you don't need to, you know, need to back up at all.", "tokens": [50514, 407, 291, 500, 380, 643, 281, 11, 291, 458, 11, 643, 281, 646, 493, 412, 439, 13, 50714], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 106, "seek": 70136, "start": 708.36, "end": 710.36, "text": " And so on.", "tokens": [50714, 400, 370, 322, 13, 50814], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 107, "seek": 70136, "start": 710.36, "end": 715.36, "text": " And so, eventually, I decided to write that thing up.", "tokens": [50814, 400, 370, 11, 4728, 11, 286, 3047, 281, 2464, 300, 551, 493, 13, 51064], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 108, "seek": 70136, "start": 715.36, "end": 718.36, "text": " Although, at my day job at that time, it was designing computers.", "tokens": [51064, 5780, 11, 412, 452, 786, 1691, 412, 300, 565, 11, 309, 390, 14685, 10807, 13, 51214], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 109, "seek": 70136, "start": 718.36, "end": 721.36, "text": " So it took me a while to get it written up,", "tokens": [51214, 407, 309, 1890, 385, 257, 1339, 281, 483, 309, 3720, 493, 11, 51364], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 110, "seek": 70136, "start": 721.36, "end": 724.36, "text": " and then it took me a while to get it through the refereeing process.", "tokens": [51364, 293, 550, 309, 1890, 385, 257, 1339, 281, 483, 309, 807, 264, 33048, 14667, 1399, 13, 51514], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 111, "seek": 70136, "start": 724.36, "end": 728.36, "text": " So it finally appeared in 1976.", "tokens": [51514, 407, 309, 2721, 8516, 294, 33978, 13, 51714], "temperature": 0.0, "avg_logprob": -0.14525422182950107, "compression_ratio": 1.6367713004484306, "no_speech_prob": 0.002346066990867257}, {"id": 112, "seek": 72836, "start": 728.36, "end": 734.36, "text": " But, you know, several things have struck me about that in retrospect.", "tokens": [50364, 583, 11, 291, 458, 11, 2940, 721, 362, 13159, 385, 466, 300, 294, 34997, 13, 50664], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 113, "seek": 72836, "start": 734.36, "end": 740.36, "text": " One, as I said before, there is this tremendous head start", "tokens": [50664, 1485, 11, 382, 286, 848, 949, 11, 456, 307, 341, 10048, 1378, 722, 50964], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 114, "seek": 72836, "start": 740.36, "end": 745.36, "text": " in working something out if you know it's true.", "tokens": [50964, 294, 1364, 746, 484, 498, 291, 458, 309, 311, 2074, 13, 51214], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 115, "seek": 72836, "start": 745.36, "end": 750.36, "text": " If you have any doubt that it's true, that really holds you back, of course.", "tokens": [51214, 759, 291, 362, 604, 6385, 300, 309, 311, 2074, 11, 300, 534, 9190, 291, 646, 11, 295, 1164, 13, 51464], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 116, "seek": 72836, "start": 750.36, "end": 753.36, "text": " If you get the idea that something is true, that's really false.", "tokens": [51464, 759, 291, 483, 264, 1558, 300, 746, 307, 2074, 11, 300, 311, 534, 7908, 13, 51614], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 117, "seek": 72836, "start": 753.36, "end": 756.36, "text": " You know, you could ruin your life for a long time.", "tokens": [51614, 509, 458, 11, 291, 727, 15514, 428, 993, 337, 257, 938, 565, 13, 51764], "temperature": 0.0, "avg_logprob": -0.0868992805480957, "compression_ratio": 1.7096774193548387, "no_speech_prob": 0.012053944170475006}, {"id": 118, "seek": 75636, "start": 756.36, "end": 763.36, "text": " And there was another thing that I saw in Peter's algorithm", "tokens": [50364, 400, 456, 390, 1071, 551, 300, 286, 1866, 294, 6508, 311, 9284, 50714], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 119, "seek": 75636, "start": 763.36, "end": 767.36, "text": " that happened to me in some other algorithms that I did,", "tokens": [50714, 300, 2011, 281, 385, 294, 512, 661, 14642, 300, 286, 630, 11, 50914], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 120, "seek": 75636, "start": 767.36, "end": 774.36, "text": " which was that you're sort of going through a process of hunting", "tokens": [50914, 597, 390, 300, 291, 434, 1333, 295, 516, 807, 257, 1399, 295, 12599, 51264], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 121, "seek": 75636, "start": 774.36, "end": 777.36, "text": " when you're trying to solve a problem.", "tokens": [51264, 562, 291, 434, 1382, 281, 5039, 257, 1154, 13, 51414], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 122, "seek": 75636, "start": 777.36, "end": 779.36, "text": " You go over there, you go over there,", "tokens": [51414, 509, 352, 670, 456, 11, 291, 352, 670, 456, 11, 51514], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 123, "seek": 75636, "start": 779.36, "end": 781.36, "text": " and you get some super structure here,", "tokens": [51514, 293, 291, 483, 512, 1687, 3877, 510, 11, 51614], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 124, "seek": 75636, "start": 781.36, "end": 783.36, "text": " and you get some data structure over here, and so on.", "tokens": [51614, 293, 291, 483, 512, 1412, 3877, 670, 510, 11, 293, 370, 322, 13, 51714], "temperature": 0.0, "avg_logprob": -0.09702309229040659, "compression_ratio": 1.8473684210526315, "no_speech_prob": 0.017157116904854774}, {"id": 125, "seek": 78336, "start": 783.36, "end": 787.36, "text": " And there was a case where I was doing priority search trees a few years later,", "tokens": [50364, 400, 456, 390, 257, 1389, 689, 286, 390, 884, 9365, 3164, 5852, 257, 1326, 924, 1780, 11, 50564], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 126, "seek": 78336, "start": 787.36, "end": 792.36, "text": " and I had come up with this really complicated super structure,", "tokens": [50564, 293, 286, 632, 808, 493, 365, 341, 534, 6179, 1687, 3877, 11, 50814], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 127, "seek": 78336, "start": 792.36, "end": 798.36, "text": " and I was convinced that it worked until John Bentley showed up in my office one day,", "tokens": [50814, 293, 286, 390, 12561, 300, 309, 2732, 1826, 2619, 43147, 4712, 493, 294, 452, 3398, 472, 786, 11, 51114], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 128, "seek": 78336, "start": 798.36, "end": 800.36, "text": " and I explained it to John.", "tokens": [51114, 293, 286, 8825, 309, 281, 2619, 13, 51214], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 129, "seek": 78336, "start": 800.36, "end": 804.36, "text": " John said, oh, no, what about this case?", "tokens": [51214, 2619, 848, 11, 1954, 11, 572, 11, 437, 466, 341, 1389, 30, 51414], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 130, "seek": 78336, "start": 804.36, "end": 806.36, "text": " So I studied it, and studied on it,", "tokens": [51414, 407, 286, 9454, 309, 11, 293, 9454, 322, 309, 11, 51514], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 131, "seek": 78336, "start": 806.36, "end": 810.36, "text": " and finally found something that would solve that case.", "tokens": [51514, 293, 2721, 1352, 746, 300, 576, 5039, 300, 1389, 13, 51714], "temperature": 0.0, "avg_logprob": -0.1728072072945389, "compression_ratio": 1.652542372881356, "no_speech_prob": 0.011847513727843761}, {"id": 132, "seek": 81036, "start": 810.36, "end": 815.36, "text": " So I was giving a talk about it at Stanford University,", "tokens": [50364, 407, 286, 390, 2902, 257, 751, 466, 309, 412, 20374, 3535, 11, 50614], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 133, "seek": 81036, "start": 815.36, "end": 819.36, "text": " and one of my colleagues from Xerox Park was along to listen to the talk,", "tokens": [50614, 293, 472, 295, 452, 7734, 490, 1783, 2032, 87, 4964, 390, 2051, 281, 2140, 281, 264, 751, 11, 50814], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 134, "seek": 81036, "start": 819.36, "end": 821.36, "text": " because he had sort of seen me struggling over this,", "tokens": [50814, 570, 415, 632, 1333, 295, 1612, 385, 9314, 670, 341, 11, 50914], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 135, "seek": 81036, "start": 821.36, "end": 824.36, "text": " but it never actually heard the explanation.", "tokens": [50914, 457, 309, 1128, 767, 2198, 264, 10835, 13, 51064], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 136, "seek": 81036, "start": 824.36, "end": 827.36, "text": " We were driving back to Xerox Park afterward, and he said,", "tokens": [51064, 492, 645, 4840, 646, 281, 1783, 2032, 87, 4964, 40411, 11, 293, 415, 848, 11, 51214], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 137, "seek": 81036, "start": 827.36, "end": 829.36, "text": " you know, I don't think I understood it really.", "tokens": [51214, 291, 458, 11, 286, 500, 380, 519, 286, 7320, 309, 534, 13, 51314], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 138, "seek": 81036, "start": 829.36, "end": 833.36, "text": " I said, oh, sorry, Howard, I tried so hard to explain it well.", "tokens": [51314, 286, 848, 11, 1954, 11, 2597, 11, 17626, 11, 286, 3031, 370, 1152, 281, 2903, 309, 731, 13, 51514], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 139, "seek": 81036, "start": 833.36, "end": 835.36, "text": " He said, you know, I understood the words.", "tokens": [51514, 634, 848, 11, 291, 458, 11, 286, 7320, 264, 2283, 13, 51614], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 140, "seek": 81036, "start": 835.36, "end": 839.36, "text": " It's just that when you came to that little trick at the end,", "tokens": [51614, 467, 311, 445, 300, 562, 291, 1361, 281, 300, 707, 4282, 412, 264, 917, 11, 51814], "temperature": 0.0, "avg_logprob": -0.12478651319231306, "compression_ratio": 1.7074829931972788, "no_speech_prob": 0.0037316849920898676}, {"id": 141, "seek": 83936, "start": 839.36, "end": 844.36, "text": " it didn't work out why that little trick didn't solve the whole problem.", "tokens": [50364, 309, 994, 380, 589, 484, 983, 300, 707, 4282, 994, 380, 5039, 264, 1379, 1154, 13, 50614], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 142, "seek": 83936, "start": 844.36, "end": 849.36, "text": " He said, oh, Howard, it's because,", "tokens": [50614, 634, 848, 11, 1954, 11, 17626, 11, 309, 311, 570, 11, 50864], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 143, "seek": 83936, "start": 849.36, "end": 855.36, "text": " you see, it's because, you're right, Howard.", "tokens": [50864, 291, 536, 11, 309, 311, 570, 11, 291, 434, 558, 11, 17626, 13, 51164], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 144, "seek": 83936, "start": 855.36, "end": 861.36, "text": " So I sort of held on to this super structure,", "tokens": [51164, 407, 286, 1333, 295, 5167, 322, 281, 341, 1687, 3877, 11, 51464], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 145, "seek": 83936, "start": 861.36, "end": 863.36, "text": " because that's how I reached the final answer,", "tokens": [51464, 570, 300, 311, 577, 286, 6488, 264, 2572, 1867, 11, 51564], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 146, "seek": 83936, "start": 863.36, "end": 867.36, "text": " even though it turned out that it wasn't necessary in the end.", "tokens": [51564, 754, 1673, 309, 3574, 484, 300, 309, 2067, 380, 4818, 294, 264, 917, 13, 51764], "temperature": 0.0, "avg_logprob": -0.16971438118581023, "compression_ratio": 1.5794871794871794, "no_speech_prob": 0.0022071748971939087}, {"id": 147, "seek": 86736, "start": 867.36, "end": 869.36, "text": " It was kind of like scaffolding on a building.", "tokens": [50364, 467, 390, 733, 295, 411, 44094, 278, 322, 257, 2390, 13, 50464], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 148, "seek": 86736, "start": 869.36, "end": 871.36, "text": " You know, you need it to build the building,", "tokens": [50464, 509, 458, 11, 291, 643, 309, 281, 1322, 264, 2390, 11, 50564], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 149, "seek": 86736, "start": 871.36, "end": 875.36, "text": " but then once the building's built, you can take it down.", "tokens": [50564, 457, 550, 1564, 264, 2390, 311, 3094, 11, 291, 393, 747, 309, 760, 13, 50764], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 150, "seek": 86736, "start": 875.36, "end": 879.36, "text": " And so that's another thing you can...", "tokens": [50764, 400, 370, 300, 311, 1071, 551, 291, 393, 485, 50964], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 151, "seek": 86736, "start": 879.36, "end": 883.36, "text": " You sometimes need a bump in the head once you've gotten to a certain point", "tokens": [50964, 509, 2171, 643, 257, 9961, 294, 264, 1378, 1564, 291, 600, 5768, 281, 257, 1629, 935, 51164], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 152, "seek": 86736, "start": 883.36, "end": 886.36, "text": " to get rid of all the stuff you've fallen in love with", "tokens": [51164, 281, 483, 3973, 295, 439, 264, 1507, 291, 600, 11547, 294, 959, 365, 51314], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 153, "seek": 86736, "start": 886.36, "end": 891.36, "text": " in the course of getting to where you're getting to it.", "tokens": [51314, 294, 264, 1164, 295, 1242, 281, 689, 291, 434, 1242, 281, 309, 13, 51564], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 154, "seek": 86736, "start": 891.36, "end": 896.36, "text": " Well, actually, the bump has gone longer than I deserve,", "tokens": [51564, 1042, 11, 767, 11, 264, 9961, 575, 2780, 2854, 813, 286, 9948, 11, 51814], "temperature": 0.0, "avg_logprob": -0.10507771928431624, "compression_ratio": 1.7419354838709677, "no_speech_prob": 0.003443208523094654}, {"id": 155, "seek": 89636, "start": 896.36, "end": 901.36, "text": " but I just wanted to make one final comment here.", "tokens": [50364, 457, 286, 445, 1415, 281, 652, 472, 2572, 2871, 510, 13, 50614], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 156, "seek": 89636, "start": 901.36, "end": 904.36, "text": " I gave it up, and I went on to do other things,", "tokens": [50614, 286, 2729, 309, 493, 11, 293, 286, 1437, 322, 281, 360, 661, 721, 11, 50764], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 157, "seek": 89636, "start": 904.36, "end": 906.36, "text": " design more hardware, and so on,", "tokens": [50764, 1715, 544, 8837, 11, 293, 370, 322, 11, 50864], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 158, "seek": 89636, "start": 906.36, "end": 910.36, "text": " so I didn't do anything more with suffix trees.", "tokens": [50864, 370, 286, 994, 380, 360, 1340, 544, 365, 3889, 970, 5852, 13, 51064], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 159, "seek": 89636, "start": 910.36, "end": 913.36, "text": " But in the early 1980s, I got a very interesting phone call,", "tokens": [51064, 583, 294, 264, 2440, 13626, 82, 11, 286, 658, 257, 588, 1880, 2593, 818, 11, 51214], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 160, "seek": 89636, "start": 913.36, "end": 915.36, "text": " and there was this guy on the other end of the line,", "tokens": [51214, 293, 456, 390, 341, 2146, 322, 264, 661, 917, 295, 264, 1622, 11, 51314], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 161, "seek": 89636, "start": 915.36, "end": 919.36, "text": " and he was talking to me about my suffix tree paper,", "tokens": [51314, 293, 415, 390, 1417, 281, 385, 466, 452, 3889, 970, 4230, 3035, 11, 51514], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 162, "seek": 89636, "start": 919.36, "end": 923.36, "text": " and it was clear that he had read the paper in detail,", "tokens": [51514, 293, 309, 390, 1850, 300, 415, 632, 1401, 264, 3035, 294, 2607, 11, 51714], "temperature": 0.0, "avg_logprob": -0.10394814483120911, "compression_ratio": 1.646090534979424, "no_speech_prob": 0.00137529952917248}, {"id": 163, "seek": 92336, "start": 923.36, "end": 926.36, "text": " but he wasn't talking like a computer scientist,", "tokens": [50364, 457, 415, 2067, 380, 1417, 411, 257, 3820, 12662, 11, 50514], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 164, "seek": 92336, "start": 926.36, "end": 928.36, "text": " and I couldn't figure it out.", "tokens": [50514, 293, 286, 2809, 380, 2573, 309, 484, 13, 50614], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 165, "seek": 92336, "start": 928.36, "end": 931.36, "text": " He was using different words,", "tokens": [50614, 634, 390, 1228, 819, 2283, 11, 50764], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 166, "seek": 92336, "start": 931.36, "end": 935.36, "text": " and so we talked for 20 minutes or so,", "tokens": [50764, 293, 370, 321, 2825, 337, 945, 2077, 420, 370, 11, 50964], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 167, "seek": 92336, "start": 935.36, "end": 937.36, "text": " and at the end of the talk, I said,", "tokens": [50964, 293, 412, 264, 917, 295, 264, 751, 11, 286, 848, 11, 51064], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 168, "seek": 92336, "start": 937.36, "end": 940.36, "text": " look, you know, I'm dying of curiosity.", "tokens": [51064, 574, 11, 291, 458, 11, 286, 478, 8639, 295, 18769, 13, 51214], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 169, "seek": 92336, "start": 940.36, "end": 943.36, "text": " You are clearly a very smart guy,", "tokens": [51214, 509, 366, 4448, 257, 588, 4069, 2146, 11, 51364], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 170, "seek": 92336, "start": 943.36, "end": 947.36, "text": " and you have clearly read my paper in a great deal of detail,", "tokens": [51364, 293, 291, 362, 4448, 1401, 452, 3035, 294, 257, 869, 2028, 295, 2607, 11, 51564], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 171, "seek": 92336, "start": 947.36, "end": 951.36, "text": " but you don't sound like a computer scientist.", "tokens": [51564, 457, 291, 500, 380, 1626, 411, 257, 3820, 12662, 13, 51764], "temperature": 0.0, "avg_logprob": -0.05115109902841074, "compression_ratio": 1.641255605381166, "no_speech_prob": 0.002863208530470729}, {"id": 172, "seek": 95136, "start": 951.36, "end": 953.36, "text": " Who are you?", "tokens": [50364, 2102, 366, 291, 30, 50464], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 173, "seek": 95136, "start": 953.36, "end": 956.36, "text": " He said, I'm a biologist.", "tokens": [50464, 634, 848, 11, 286, 478, 257, 3228, 9201, 13, 50614], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 174, "seek": 95136, "start": 959.36, "end": 961.36, "text": " So thank you very much.", "tokens": [50764, 407, 1309, 291, 588, 709, 13, 50864], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 175, "seek": 95136, "start": 966.36, "end": 969.36, "text": " Any questions?", "tokens": [51114, 2639, 1651, 30, 51264], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 176, "seek": 95136, "start": 969.36, "end": 972.36, "text": " I want to know what B stands for in B3.", "tokens": [51264, 286, 528, 281, 458, 437, 363, 7382, 337, 294, 363, 18, 13, 51414], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 177, "seek": 95136, "start": 972.36, "end": 975.36, "text": " Everybody does.", "tokens": [51414, 7646, 775, 13, 51564], "temperature": 0.0, "avg_logprob": -0.17901166280110678, "compression_ratio": 1.1367521367521367, "no_speech_prob": 0.0071327234618365765}, {"id": 178, "seek": 97536, "start": 975.36, "end": 980.36, "text": " So, you just have no idea", "tokens": [50364, 407, 11, 291, 445, 362, 572, 1558, 50614], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 179, "seek": 97536, "start": 980.36, "end": 984.36, "text": " what a lunchtime conversation can turn into.", "tokens": [50614, 437, 257, 6349, 3766, 3761, 393, 1261, 666, 13, 50814], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 180, "seek": 97536, "start": 984.36, "end": 986.36, "text": " So there we were.", "tokens": [50814, 407, 456, 321, 645, 13, 50914], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 181, "seek": 97536, "start": 986.36, "end": 988.36, "text": " Merida and I had lunch.", "tokens": [50914, 6124, 2887, 293, 286, 632, 6349, 13, 51014], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 182, "seek": 97536, "start": 988.36, "end": 990.36, "text": " We had to give a thing a name,", "tokens": [51014, 492, 632, 281, 976, 257, 551, 257, 1315, 11, 51114], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 183, "seek": 97536, "start": 990.36, "end": 997.36, "text": " and we were, so B, we were thinking B is, you know,", "tokens": [51114, 293, 321, 645, 11, 370, 363, 11, 321, 645, 1953, 363, 307, 11, 291, 458, 11, 51464], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 184, "seek": 97536, "start": 997.36, "end": 999.36, "text": " we were working for Boeing at the time,", "tokens": [51464, 321, 645, 1364, 337, 30831, 412, 264, 565, 11, 51564], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 185, "seek": 97536, "start": 999.36, "end": 1002.36, "text": " but we couldn't use the name without talking to the lawyers.", "tokens": [51564, 457, 321, 2809, 380, 764, 264, 1315, 1553, 1417, 281, 264, 16219, 13, 51714], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 186, "seek": 97536, "start": 1002.36, "end": 1004.36, "text": " So there's a B.", "tokens": [51714, 407, 456, 311, 257, 363, 13, 51814], "temperature": 0.0, "avg_logprob": -0.20588682193567256, "compression_ratio": 1.6, "no_speech_prob": 0.008810681290924549}, {"id": 187, "seek": 100436, "start": 1004.36, "end": 1007.36, "text": " It has to do with balance.", "tokens": [50364, 467, 575, 281, 360, 365, 4772, 13, 50514], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 188, "seek": 100436, "start": 1007.36, "end": 1009.36, "text": " There's another B.", "tokens": [50514, 821, 311, 1071, 363, 13, 50614], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 189, "seek": 100436, "start": 1009.36, "end": 1011.36, "text": " Rudy was the senior author.", "tokens": [50614, 38690, 390, 264, 7965, 3793, 13, 50714], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 190, "seek": 100436, "start": 1011.36, "end": 1013.36, "text": " Rudy was several years older than I am.", "tokens": [50714, 38690, 390, 2940, 924, 4906, 813, 286, 669, 13, 50814], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 191, "seek": 100436, "start": 1013.36, "end": 1017.36, "text": " It had been, you know, I had many more publications than I did.", "tokens": [50814, 467, 632, 668, 11, 291, 458, 11, 286, 632, 867, 544, 25618, 813, 286, 630, 13, 51014], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 192, "seek": 100436, "start": 1017.36, "end": 1019.36, "text": " So there's another B.", "tokens": [51014, 407, 456, 311, 1071, 363, 13, 51114], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 193, "seek": 100436, "start": 1019.36, "end": 1023.36, "text": " And so at the lunchtime, we never did resolve", "tokens": [51114, 400, 370, 412, 264, 6349, 3766, 11, 321, 1128, 630, 14151, 51314], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 194, "seek": 100436, "start": 1023.36, "end": 1026.3600000000001, "text": " whether there was one of those", "tokens": [51314, 1968, 456, 390, 472, 295, 729, 51464], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 195, "seek": 100436, "start": 1026.3600000000001, "end": 1031.3600000000001, "text": " that made more sense than the rest.", "tokens": [51464, 300, 1027, 544, 2020, 813, 264, 1472, 13, 51714], "temperature": 0.0, "avg_logprob": -0.11863329055461477, "compression_ratio": 1.5918367346938775, "no_speech_prob": 0.001946610165759921}, {"id": 196, "seek": 103136, "start": 1031.36, "end": 1033.36, "text": " So what he likes to say is,", "tokens": [50364, 407, 437, 415, 5902, 281, 584, 307, 11, 50464], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}, {"id": 197, "seek": 103136, "start": 1033.36, "end": 1038.36, "text": " the more you think about what the B in B-trees means,", "tokens": [50464, 264, 544, 291, 519, 466, 437, 264, 363, 294, 363, 12, 3599, 279, 1355, 11, 50714], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}, {"id": 198, "seek": 103136, "start": 1038.36, "end": 1041.36, "text": " the better you understand B-trees.", "tokens": [50714, 264, 1101, 291, 1223, 363, 12, 3599, 279, 13, 50864], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}, {"id": 199, "seek": 103136, "start": 1044.36, "end": 1045.36, "text": " Fair enough.", "tokens": [51014, 12157, 1547, 13, 51064], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}, {"id": 200, "seek": 103136, "start": 1045.36, "end": 1047.36, "text": " Any other questions?", "tokens": [51064, 2639, 661, 1651, 30, 51164], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}, {"id": 201, "seek": 103136, "start": 1047.36, "end": 1049.36, "text": " Let's thank the speaker.", "tokens": [51164, 961, 311, 1309, 264, 8145, 13, 51264], "temperature": 0.0, "avg_logprob": -0.26945262119687835, "compression_ratio": 1.3257575757575757, "no_speech_prob": 0.02865605801343918}], "language": "en"}