ロスではなくラスベガスで執筆活動
こちらの時間は深夜 2:30 PST 12/04 (木) 。
ラスベガスです。
最初に投稿した記事で「ロスで執筆活動」と書いていましたが
もはや睡眠時間が足りな過ぎて一部自分が壊れていたこと判明。
re:Invent に参加してから毎日歩くは歩くし（別記事にて
セッションの内容が敷居が高いしで壊れそうでした
特に初日の bootcamp（ビリー隊長が兵隊を集めて教えるアレ
数時間講義を聞いた時点でこれは駄目だと痛感。
そうです。私みたいな日常会話すらまともにできない人が
技術セッション内容を丸腰で聞いても理解できないのです。
使うとしたらアレ。ほんやくコンニャク（違う
対象セッション
A Special Closing Keynote with Dr. Werner Vogels (KEY005)
これっきゃない。
ワーナーさんのセッションが日本人向けのラップアップと衝突しているので
どちらに参加するか悩みました。
ただモニタ経由でみるのであれば、現地にいった利点がなくて日本からでも
同じ内容が見えてしまうし（その方が時差の問題もなく快適な噂もあるけど
やっぱり。
会場の熱気を感じたいし、ワーナーさんの情熱をその場で受け取りたいから
re:Invent に参加したはず。
その中でいちばんの課題は英語。youtube ではないから日本語の字幕など
どこにも出ません。英語の字幕はスクリーン横に表示されていますが
それから読み取ることができる人は、もはや英会話に困らない人では。
選択したツール
初日に痛感したのは翻訳アプリを使えば全て解決するものでもないこと
・同時翻訳できるツールを見つめていると話者を見ないことになるので
討論には不向き
同じテーブルのメンバから「あいつら翻訳している間に時間かかるから
日本語に翻訳したテキストとかないの？」ってスピーカに相談いただく
同じチームにいたアメリカンの優しさと悲しさに触れる
・会場に表示される字幕や文字起こしツールで英文を記録したとしても
内容を理解しようとすると時間が経過してしまう
スクリーンならまだしもPCやモニタに表示される英語はとにかく小さい
老眼の敵です。
小さくて読めないし、読めても理解が進まない悲しみよ、こんにちは（誰
画面の文字が見えないのであれば、あと残されるのは聴覚と嗅覚だけ
匂いで嗅ぎ分けることができない場合にはやっぱり同時通訳ですよね。
自動翻訳マシーン
iPhoneとAirPodsを同時に利用することで、ほんやくコンニャクの
イヤホン版になるとの噂を聞いておりました。
ただ私は中華携帯＋中華イヤホンの派閥なのでそんなことはできず。
自分の持ち物だけで可能な範囲でワーナーさんの発言を逃さず聞いて
持ち帰ることを最大の目的としています。
対策グッズとして現地持ち込みした携帯2個とPC1台で以下を分業。
| 名前 | OS | 用途 |
|---|---|---|
| 翻訳機 | 不明 | 会場で配布されたハイテクマシーン |
| 1号 | android | Google同時翻訳/会話モード |
| 2号 | iPhone | 同時通訳pro |
| PC | Windows | LINE works aiNote |
翻訳機はもろ本丸。一個前のkeynoteでも利用しておりましたが
やっぱり日本語が聞こえてくるのは安心です。
携帯1号は会話モードを起動しっぱなしにすることで
英語で会話された内容を文字でみれるようにしておくことが可能
携帯2号は同時翻訳されるため、英語ではなく日本語が流れます
ただ一定以上の文字数になるとボタンを押下しないと続かない
PCのが今回の目玉。あえて otter ai ではなく LINE works さんの
aiNote を採用したのは理由が
※ 単純に巷で推される人気者にはのっからない派閥なだけな噂も。
（あとで画像を貼り付ける）
罰ゲームスタート
待ち時間を流れる壮大なクラッシックが完了したあと、盛大な拍手と共に
ワーナーさんが登場。そして自動翻訳機から流れるような日本語が・・・。
ていうか
英語ですら何も聞こえない・・・笑
電池切れた機材を配布するのはやめていただけませんか。
この時点で選択肢、生声のみに確定。
途中記載中
成果物
ひとまず成功した気がするので本日 keynote投稿しておきます。
みなさんは
気持ちが悪いので具体的な記載や日本語訳は別途・・笑
成功の秘密は別途ｗ（駄目
Evolution of Software Development and Developer Qualities （長すぎるので折り畳み）
参加者 1 06:23
If you go down like you're going to go over to the ROM you know how you block down what goes over the wind but if you keep going straight, it'll take you over to the Caesar form and then going out to the back door of the Caesar forum.
参加者 1 07:02
And then if you want to stop and see the EWS Tech house, we've got the drone.
No, that's over at the Caesar forum
参加者 2 20:18
of the development. I've heard that before.
参加者 3 21:33
Oh, you know, organizing yesterday's program cards which got dropped figuring out the machine keeps stopping on part 442 in this stilly awkward bug with the bala compiler and talking to you anything new at head office
参加者 3 21:58
Yes, that's the new higher language.
参加者 4 22:00
It's amazing now anyone can write code
参加者 3 22:05
I don't know about anyone.
Writing software is pretty tricky.
参加者 3 22:14
Software
参加者 4 22:53
visual programming code's history marting drag and drops The future dragon drop is still cold It still has bugs and it still needs
参加者 5 23:40
Is this for real? Could cloud eye in there less engineers
参加者 6 23:57
servers in minutes, scale on demand pay as you go.
参加者 4 24:51
Is this the end of development as we know it
参加者 2 25:53
for every day for a something view, an open mind for a different view and nothing else matters.
Good morning hello hello guys Take a video shows to every generation of developers that faced a new wave of change.
Tools evolved architects evolve expectations evolve.
参加者 2 26:38
I've been given this kiner since 2012 and I've got all of them as our long T shirts.
By the way,
参加者 2 26:59
I have still things to do.
I'm not leaving Amazon or anything like that but I think that after 14 or three events, you guys, our own young, fresh new voices There are so many amazing engineers at Amazon that have great stories to tell, to teach you, to help you, to educate you and I think it's time for those younger different voices of AWS to be in front of you
参加者 2 27:34
but I'm only for sweeten.
I'm not giving UPS off or anything like this.
This is my decision to make sure that you get to hear different voices and just mine.
Now let's talk about the other elephants in the room.
I've visited ADBS posts all over the world and at the very reason these days, one impression that keeps coming home
参加者 2 28:30
some scales will become obsolete new ones may ignorage.
So maybe we should rehrase and rehrase this question.
Yeah, will AI make me obsolete?
Absolutely not
参加者 2 29:01
if you evolve Yeah. And if I look at the past years, an Amazon where though we've been using all of these new tools we've seen how you can evolve over time and still be a great engineer with just remusion of tools in your hands because we evolved
参加者 2 29:35
and so change is constant and this has always been the case.
It's nothing new. Yeah let's go back in time a little bit just for that.
When I went to school, I was taught 68th class of December
参加者 2 30:02
and in the 60s, you got suddenly got compilers and you didn't really remember anymore what kind of assembly those compilers spit out
参加者 2 30:22
actually translated into machine code.
And so that wasn't impul to me.
参加者 2 30:46
Thirdly, they supported the shift to help us a clearer control water flow and helped them escape the chaos of construction to
参加者 2 31:16
and in the late 1990s, I was almost still working as a modernist and in 98, such a favorite moment the girls accelerated to such a point that the thing became brazing off pieces of these monuments into two circles and each of us had ownership of the service, had their own interface and it completely changed how developers worked moved faster they got independent of other teams.
They owned their assistance and tends.
参加者 2 31:54
And over time, the industry at large became adopting these kinds of practices as a practical Model for building and, and operating large scale distributed systems.
Actually what about the JWS in those days?
But in the 2000, most developers, they were still building and deploying things on the premise.
参加者 2 32:33
Developers cert only had know the amount of infrastructure freedom to pargrams not waiting for a heartbreak.
And this required new skills and developed somewhere I don't give work to do worldware.
参加者 2 33:37
And today's environment occurs a GLE and that's the new work for, is there a need to do work for them next year?
Five years from now 10 years from now of course there is.
Yeah
参加者 2 34:03
they already saw examples of developers becoming dramatically more productive with AI assistant workflows.
But none of this removes the work that only you can do.
Remember the work is yours, not that of the tools.
It is your work that matters.
Our Jews have changed so many times over the course of my career that they will continue to change.
We're still builders we're still important.
Nothing has changed there.
参加者 2 34:52
There's never been a time to be more excited about being a developer bas Not that long ago in the interview, he talked about this as that we're living at the epicenter of a H multiple simultaneous golden Ages cooked together a spacecraft or artificial intelligence projects.
参加者 2 35:24
It is how these breakthroughs actually reinforce each other.
Progress in one field accelerates progress in the other fields.
And this actually made me think back in history at the time
参加者 2 35:54
okay after a period of darkness, the dark ages, the Middle Ages
参加者 2 36:12
but the Rens was a period where everything changed because people became curious.
Curiosity was absolutely exposing and the result of that is amazing scientists and philosophers.
And if you look at this, the medicy is probably the first version of the fan capital.
参加者 2 37:07
Because of curiosity, a bentil was invented.
That seems nothing to be invented today but there was a major field, the fact that they start thinking about something called a vanishing point.
Well, certainly that if you compare paintings and drawings to before the Renaissance, they were all flat.
参加者 2 37:41
And then these schools like the microscope and the telescope, of course invented by those people.
I'll say anything and then the printing quest of course, we all see as sort of the pinnacle of invention in the Rediss.
But I was not just my invention.
A good word actually used the wine press as his first state but that was not the only thing he needed to invent.
He needed to invent prove all time
参加者 2 38:17
basically that you have characters that you can put together you have to invented ink that actually could be put on characters.
参加者 2 38:35
It was a time where art and science were part of the same pization.
Creative and technology evolved together.
I'll spend some time to think about what might people so effect in that world they were curious, they questioned assumptions they learned broadly and applied that learning deeply.
They didn't see boundaries between fields.
They built bridges between them.
They were also both experimenters.
They sketched, they measured they failed they tried again.
They learned by doing
参加者 2 39:25
I think by putting them together I think we are again in the time of Renaissance.
A you
参加者 2 39:42
of those scientists in the Renaissance are just as relevant today.
So BOT C framework than are all the Renaissance developed but I want to show you today this framework hopefully help you evolve and be successful in this new era as well.
What is crucial and all of this is the first quality that you need to nurture is you need to be curious.
Curiosity is critical. As developers, you always had to continuously learn because everything changed all the time
参加者 2 40:32
and every developer I've met has this instinct to take something apart and look at how it works.
And it's also what drives us here the desire to understand, to improve, to build.
You have to protect that instinct.
Stay curious because curiosity leads to learning and invention equally important for learning is two things.
For all new inventions you need to experiment and to experiment.
Well, you need to be willing to fail after all din.
参加者 2 41:29
And by the way, an experiment is not an experiment.
If you already know the outcome it drives experimentation, drives the learning and this willingness to fail is crucial.
When I learned a new language that THEC for English and port GS and German I find that the same principles apply the best way to learn.
参加者 2 42:06
But really leading begins be you stumble into a conversation and someone helps you to get grants, software.
It works the same way you can read documentation any but it is to failed to build.
But the broken assumptions that really teach you how a system behaves.
参加者 2 42:41
And there are certain things that you can only do and only learn by doing reading, watching listening only takes you so far.
But green learning happens when you engage when you so pressure, when the outcome matters there is a relationship between stress and performance called the jerks and law think she's a bilk too little pressure, you disengaged too much pressure and you're overwhelmed.
The sweet spot is so well that rising slow where Curi meets challenge
参加者 2 43:36
you have to put yourself in positions that test you.
Now there's a whole story behind this and that newspaper that you all found in your seats today The Kournal.
There's an article in there by Andy Moorfield who writes about this
参加者 2 44:06
learning. There's another side to it.
Learning is SOC we're not here.
Al to sit in a room and listen to one person is telling you exactly what you're doing.
The thing you really learn is by talking to a child.
参加者 2 44:40
I have coffee with a friend and top of our systems one of the best ways to stay sharp is to be around a lot of people who are building things.
And for me, that often happens among the world.
I travel a lot for the world and those trips keep me connected to how people are actually using technology, not just how we imagine they might be.
This year I was very fortunate.
I took two month long trips to Sahara Africa and the other to Latin America.
参加者 2 45:34
This is actually also it took me 20 more years to actually end up on the Amazon.
And so what you is a beverage company they support communities along the Anson River in a way that they give human people in the economic future so that they don't leave their villages to go to the big towns.
It is a brilliant experience and a great example of how doing good can be profitable.
At the same time, the Amazon level was beautiful
参加者 2 46:18
but it reminded of something else that I learned that not all of this is so great
参加者 2 46:43
from just a thousandth of the world's three million years to rid the oceans of plastic, they need not only to clean up what there's already out there but also to stop new wests from entering the ocean via these rivers.
And they do that. They've created a river body using drugs, AI camera analysis, even GPS.
参加者 2 47:23
The Amazon is not a big loser at all but this computational Model that makes pals, these AI cameras that are bridges that are all pack of ships
参加者 2 47:45
and helping them position their clean objects for a maximum impact.
Now another thing that
参加者 2 48:04
this is the headquarters of the Ministry of Health and this is the Health Intelligence center in the evacuation centers, huge screens displayed near real time data from 40S peers of health care facilities across the country.
They've built in such a system that ingests and processes vast amount of healthcare data visualizing everything from disease outbreaks to maternal health outcomes.
And they use it to create new Poss.
They've cleaned this Model which shows that which parts of the country
参加者 2 48:54
I'm more than a 30 minute walk away from a health care provider and they use this data to strategically place new maternal health facilities in all areas.
They use data to drive policy and to actually get them not visits that actually I mean, most of these trips I get blowing away every time especially about how young companies are getting some of the world's hardest problems.
参加者 2 49:39
People will just bow a dollar or two dollars in the morning, buy some goods please tell on the market or on the streets and in the evening they will give these dollars back and hopefully have 40 or 50 cents which they need that pay that's enough to go buy some food but it is not enough though, to also buy the gas.
参加者 2 50:18
This is absolutely beautiful.
So this is your company of efforts.
They came up with a completely different solution.
They basically built these kind of ATM machines with ethanol in them and this whole chemistry that people have and they can basically walk up to the EFT to the machine, plug in that chemistry and ask for five cents of tax, which will be enough to cook their food, their hutch.
This is what happens when developers apply their skills to real human chechnology.
参加者 2 51:05
You have built the foundations of the digital world and today you are the ones journeying through to a, from possibility into something useful safe, scalable.
The developers like you were essential in the past they're essential today and you will be essential in the future.
The United Nations expects that by 2050 we have two billion more people on this planet.
I was going to feed them.
参加者 2 51:40
I'm going to make sure we have an economic future I'm going to make sure we have health care that's on us to develop technologies so that it can help solve some of the world's biggest problems as technologists have that ability to do.
And if you look at some of the Jew who spent so much of their time not just to actually build some things in the Model of your room, but actually help others the aim of the USS.
参加者 2 52:41
There are 265 heroes across 58 countries but we're constantly amazing.
参加者 2 53:14
Graphy absolutely embodies the raises development.
He doesn't just micros he builds communities.
参加者 2 53:44
So the first quality I think that a renaissance in the real earth needs to have is to be curious.
And I like this quote from Golf that we are not what we know, but what we are willing to learn.
Another quality that I think renaissance developer has is that he thinks in systems mean just come with me for a moment if you don't really understand, I don't mean the computer system in this case, but in the system.
So
参加者 2 54:25
in the 70s, an ecologist called Dnella Medows began studying how complex systems behave.
It's harder to computer scientists but her insights describe our world of software perfectly and she writes a system is a set of things, peopleselves or whatever interconnected in such a way that they produced no wild patterns of behavior.
参加者 2 55:03
Let me give you an example by the way that is not compute.
One of the most striking examples of systems comes from ecology in the early 20th century the rules were redced from Yellowstone National class.
参加者 2 55:37
When we re induced in 20,010, Bols back into the top slowly, the park started to move.
The vegetation returned to reavers came back.
Even rivers changed course.
The most didn't move the rivers they didn't they changed the behavior of the overall system.
That single feedback predator and cave we shape the balance of the entire system.
And as structure changes, behavior changes and that feedback changes, outcome changes.
That's what's called systems figure.
So
参加者 2 56:25
then thinking in systems is thinking in complete systems not just the United States parts, every surface, every APR, every cue is part of a larger system.
You can't change one part in isolation
参加者 2 57:01
it creates new patterns.
Some stable, some not every dynamic system is shaped by feedback models.
Real fortunes, sometimes called positive Ns, amplify change.
Balancing loop is loose, counteract the change and push the system back into an equilibrium.
Melan thoughts that once you see patterns like this, you start to see their small, well placed changes might shift the overall systems we are L.
参加者 2 57:39
Then I wrote a paper FV papers called Lage Call it Basis to intervening system which put all these things together some words that we know from computer science on a daily basis positive and negative feedback groups.
参加者 2 58:02
Take a picture of the QR code and that's your homework for the coping the Renaissance developer things existed.
And to build resilient systems, you need to understand the bigger picture.
参加者 2 58:36
And if you step into this board of field, you realize that the ability to express your thinking clearly is just as critical.
And this is why I believe that one of the most important things for an engineer or a technical be can do for the engineer is to practice developed strong communication skills.
Let me give you a temple. Let's go back two years that we hit the THU the architect.
参加者 2 59:07
I don't know if you remember this picture.
This is the Amazon host base and I explained to you how we have defied hop base
参加者 2 59:49
to be able to do this. It is important not just for us as engineers but as communication tools towards the business.
参加者 2 1:00:33
Describe your system and the capabilities and the opportunities to risk the business communication is crucial.
Human languages it's a bit of a gumus, isn't it?
Although that was natural language it's ambiguous but we have so many different senses at the same time that we can turn this natural language into something less ambiguous.
参加者 2 1:01:36
We can precisely indicate in the machine what we wanted it to do.
参加者 2 1:01:50
We increasingly communicate with in natural language which is ambiguous.
So we need to help to develop mechanisms to reduce the ambiguity of that language and reduce the ambiguity of the human
参加者 2 1:02:19
I'm G and our history's life stories, perspective development thanks to our structured programming environment
参加者 2 1:02:39
relying on mediculous specifications that guided as long as 4 to 5000 lines of cold a blue.
参加者 3 1:03:14
Sometime last year, I started noticing that as I was vibrating more and more, I had a communication problem.
I was spending more and more time trying to describe to the AI what I wanted to do.
The code that the AI generated was good but the end soccer didn't do what I wanted it to do.
I would often try to start over with a new pop to start all over again and again.
I noticed that over time I wrote longer and longer, more detailed prompts trying to define what the software should do.
参加者 3 1:03:49
I would write these long mark details prompts in of Sitian and mark down and then I would copy a piece it over to my quoting assistant.
I was essentially creating a specification to better communicate with the A's that I wanted to.
Software specifications can clearly communicate how a system should behave, what it should.
And like Berner has said, many systems in history have been based on specifications like Apollo missions.
参加者 3 1:04:20
It felt like specifications were exactly what was missing from my interactions with my coding assistant.
Then I thought, how can we use this idea of specs as prompts?
What would spec driven development look like?
And this spark of an idea led us to start work on the hero IDE with this idea though we now had a new communication challenge.
How can we communicate and validate this idea with potential users?
We didn't know what it would look like and we didn't know what users wanted.
参加者 3 1:04:55
It's also kind of difficult to describe what spect driven development PL they hadn't seated before and we hadn't built it yet.
One of the best ways to get your ideas across is to quickly build a prototype, something that your users can see and touch.
Rapid bototyping is of course not a new idea.
Let's go back to the 60s, to the time of punch cards.
In 1964, Delf Egalara SRI had a rough idea of a device that had wheels on the bottom.
参加者 3 1:05:28
You would slide it across the tabletop to point to something on a computer screen.
And we can probably all picture this device in our head.
But can you imagine going back to the 60s and trying to describe what a mouse is to this guy?
Angelmar's team rapidly built a prototype out about the block of foot with a wooden shell and a wheel on the bottom and a button on top.
And this rough I communicated better but a mousewise better than any drawing or description could have
参加者 3 1:06:01
it. Let people put their hands on it and get the concept immediately.
It had great economics it felt great in their hands and like the mouse, Rapicars idea was crucial for CIA.
We knew that the economics of spectrum and development was going to be important how it felt in the hands of the developed world.
We got the radical prototypes for how we thought spectrum development could work and we put those prototypes in the hands of some internal users and we asked them to use CURA every day.
参加者 3 1:06:33
That was going to be the best way for us to understand what felt good and what didn't feel good.
This is a great example of what AI can do for us.
Now AI has fundamentally enabled graphic prototyping for software.
I'm sure that in the past we've all spent months manually putting a single idea and now we can give our users real working prototypes in minutes to give feedback about what feels like
参加者 3 1:07:03
rapidly prototyping Ciro.
Even let us use Kiro to build Kiro.
Our very first prototype of the ciro IDE could only do fine coding.
But from that moment on, we were able to use the CUR IDE to generate the code for the cure and by using the CURA IDP to build out the product, we were able to iterate through many items of how spective in development could work.
One idea was to test from the development steps, taking inspiration and CVD techniques.
You would describe the change that you wanted
参加者 3 1:07:40
and Kiera would generate tests to validate the behavior that you described.
Kiira would then generate application code and made sure that it passed those tests.
But we found that developers couldn't always capture the nuance of what they wanted their software to do or look like with tests only.
We tried out traditional technical specifications similar to those that Berner described for the Apollo guiding system.
参加者 3 1:08:07
They describe the system as a whole and each component in detail.
With this, you'd add a new feature to the overall spectrum and KERA would kick off voting tasks based on the changes that you've described.
These facts are great context for the AI and even for developers who aren't very familiar with the Spok.
But for real world projects, they sometimes became overwhelmingly long.
So it was difficult to figure out where in this long spect you should put your changes for your future.
参加者 3 1:08:39
We took a step back we looked at how we already worked as a team.
We had a new feature idea we would describe it we would work through product requirements, review a design dog and create print tasks and we thought that perhaps specks could follow this same pattern and that became feature driven specs.
We separated the flow into three documents, requirements, designs and tasks and we found that this gave developers more control over the AI and let them better communicate when they wanted to.
参加者 3 1:09:14
All of these rapid prototypes gave us critical user feedback that led to today's spectrum and development workflow that's in the CIRO ID with a SPECLC now in place in Curo, we could use it more effectively to keep building up the products because now we can communicate more precisely with the AI what we wanted.
Often with vice poting I find that we give these very ambiguous tasks to the AI.
We might say build me a WR
参加者 3 1:09:45
and out of this fury short prompt, there's probably a million possible different CL but probably only one of those is what you have in your head.
With five voting the AI is going to take its best guess as to what meant those genery CO but now leaves you to iterate on the code with the AI instead of on what originally meant with spectrum and development have an opportunity to refine what you need more precisely through Spec.
You can give the curo IDE that same ambiguous task
参加者 3 1:10:19
but instead of jumping into the code right way, it's going to first generate requirements, designs, task and if those don't match what's in your head, you have the opportunity to ask Europe to change it or find it.
Let's walk through a real example of a production feature that would be built on the CURRO ID using spectrum development system notifications.
参加者 3 1:10:43
We started out with a little annoyance that we ourselves experienced with an early version of hero identity.
Agents can take a while to complete their coding work and meanwhile we would switch away to a different hou but then the agent would be user input or it would complete its work and we would have no idea that it was sitting there IDLE while we were away doing other work.
参加者 3 1:11:07
So we set out to build a feature that would notify the user with the agent needed his attention.
We started by having CUR generate a stat and the generator requirements actually helped us to think through some details like which agent action should trigger notification.
When we got to the design phase, we'd expected this to be a pretty simple integration into the CURO agent code.
参加者 3 1:11:32
But instead KIA generated this very complex design that would build an entirely new notification system directly in our agent company.
Now if we had lived food at this, we would have ended up with a lot of hope that we didn't actually want.
But instead, the respect process helped us quickly realize this was a much bigger project than we originally thought.
参加者 3 1:11:56
We iterated on the spec to first focus on building a notification system directly outside of the agent code and directly in the underlying ID code.
This needed to be built on top of the electron's native notification.
E KI is based on code Oss, which is A, it's coves that spans 10 years of development and two million likes of Co.
It can be difficult for any developer to figure out where they need to make changes in this goods.
参加者 3 1:12:27
The CURA spec workflow actually helps us to explore and understand where these changes need to be made.
And once these changes were in place, we're again able to use spectrum development to integrate it into our agent code.
Throughout this project, we were able to quickly iterate on our steps to that exactly what we wanted from the with spectrum and development we would have shipped.
We shipped this agent roughly half the time as if we had fled it.
参加者 3 1:12:58
In our experience building Karo, we realized that ANI has changed how we communicate and how fast we can iterate.
We can iterate on the software design by communicating with the AI through steps and we can iterate on what the software does by putting real working applications fast into our user's hands to get feedback.
AI inspect driven development helped us to build a better cure I and this was in large part due to more precise communication
参加者 3 1:13:32
with their users, rapid prototypes and with AI and specs, natural language doesn't have to mean a high ambiguity.
And personally, this is what I think makes Kiera IDE special.
Kiera IDE brings precision to natural language and with that
参加者 2 1:14:02
communication is so important as player shows, it leads to systems that have fewer mistakes actually.
So let me tell you a story again sort of about my own view.
When I went to computer Science school, there was a class called interview Technique and I thought, wow, interviewed because feed by journalists and fits my L.
No it was how to learn to talk to your customer to really try to understand what he or she actually
参加者 2 1:14:40
really want because they may talk to you or whether there's a technology solution.
I'm doing anything about technology wouldn't be the first time these days Feb I mean a customer who says what should I be doing with JDI and then because really trying to figure out I'm fairly apologize to answer your question with a question.
But why are you asking me this?
Most of this is fear of missing out and so really diving into
参加者 2 1:15:16
the customer to understand what's the problem they want to solve what's the opportunity that they see all of that it's communication that we as engineers should have.
Let's come to the 4th what I call the 4th quality of the Renaissance development.
D the developer is an owner ownership.
I've spoken about it before, but today I want to focus on one part of it.
Only the quality of your software.
AI bel builds face bigger systems, explore more ideas, move faster than ever before.
参加者 2 1:16:04
This news will, as the modern day philosopher said, helps us build harder, better, faster, and stronger.
And if we use this stof correctly, they can help us reduce high quality software.
But there is a risk to, you know how some developers are starting to use these terms.
Fine coating is fine but only if you pay close attention to what is being built.
We can't just put a level on your IDE and hope that something good comes out.
参加者 2 1:16:43
That's not sofer engineering that's gambling and you need to be out there somewhere.
Remember what I said at the beginning the work is yours, not the tools.
If you subject to regulatory requirements, say health care, financial services elev if AI creates a code that fderly violates the regulatory requirement, you don't go to the regulation and say a, it's then your responsibility.
参加者 2 1:17:30
You're almost changing.
You will write lest go because generation is so fast you will review more codes because understanding it takes time and maybe you like the code yourselves.
Comprehension comes with an act of creation.
When the machine works it, you will have to rebuild that comprehension
参加者 2 1:18:05
and it's one of the two main challenges that I hear when I speak with developers about this new style of har A ion generate farther that you can understand that
参加者 2 1:18:23
allowed Sofa to move towards production before any lung has truly validated what it actually does.
The second challenge, of course is how in Asia Player showed a perfect example of that.
Earlier the Model produced and designed a P that was completely wrong for the architecture.
参加者 2 1:18:54
Sometimes they propose solutions that are completely inappropriate or over engineered or ignore your sister's own pals.
These outcomes look plausible, but they're not grounded.
In reality I think we're making progress there.
We're developing practices that are expected of development which play showed how can dramatically improve quality children.
Cara can actually use a reasoning with your specification to create code that is verified.
参加者 2 1:19:35
We showed how we can also use a lot of obesity to ensure that Drogs generated against AWS Apis is correct.
We also see many developers looking at their CSCD pipelines through building more and more automated testing.
These are all types of mechanisms
参加者 2 1:20:20
about 10 measures. In the early days of Azog ' executives were required each year to spend two days with custom reserves and 10 calls from customers such that we would July understand what the customers were going through and not just a slowly executes Jeft himself to
参加者 2 1:20:59
the customer agent says is going to return the table.
Then indeed the customer returns also returned the table because this tapps call us a gentleman's at the age he says how did you know that?
She says, well, 70 percent of those tables are coming back.
参加者 2 1:21:36
But without a mechanism, nothing changed because everybody already has good intentions.
So he introduced a new mechanism, Amazon's version of Toyota's famous Endum Ch.
So the enden court in Toyota was the principles no car should leave the production line with a no defect and that any engineer would go blind, pull these chords and bring the whole manufacturing life to the standstill until the defect was fixed.
参加者 2 1:22:12
And that the customer service agent thought it was important to make the product inviable.
That made alarms go off if the people that were responsible for the policy GT fixed.
But before this everybody already had good intentions.
But until we introduced the mechanism, nothing changed.
参加者 2 1:22:47
Each team builds their own wonders which the scale the major of their work.
The SN team, for example, uses something which they call durability reviews but an engineering policy change that might touch durability A C and Model the risks they write down, the changes they listed.
Every srapid megal let out the guard grails to keep the data safe in this mechanism is a fairly powerful effect.
It turns durability from a property of cult
参加者 2 1:23:25
into the habits of the organization.
It makes engineers think INF filiar parts not happy paths and it shows why mechanisms matter.
They convert good intentions into consistent outcomes.
参加者 2 1:23:56
It's like being 12 years old and standing in front of the class but they're a very important meganism because they create a moment where intent and implementation meet, where another engineer can cush in a substance.
参加者 2 1:24:33
It goes to control point to restore balance.
It is where we bring human judgment back into the loop and make sure that the software actually does what we expect it to do.
I encourage all of you to continue and you increase your human to human coach with you.
When senior Engunier engineers work FOK out together, it becomes one of the most effective learning mechanisms.
We have
参加者 2 1:25:17
genius, fresh eyes, wasmall details others look this is how we transfer knowledge and how do we grow the next generation of buildings AI will change many things but the crust is still lived person to person
参加者 2 1:25:51
you own it. The last quality that I'll talk to you about is that I think
参加者 2 1:26:10
I have nothing to do with maths, stop mathematics or maths or whatever you want to call it.
Actually,
参加者 2 1:26:35
it's about having deep domain experience with all who have knowledge that spans many different subjects.
Davinciin probably was the absolute best example of a polymas because he worked across so many different discipline.
He was a painter, he was an engineer, he was an anconist and he was an inventor.
I did not expect you all to become a da Victor but you should expend your knowledge.
参加者 2 1:27:28
I'm really highly specialized in one area.
I'm telling you I guess an interesting story out of my own experience.
This is my old mentor and friend Jim Grave.
参加者 2 1:27:54
Every transaction you do today, you have gym to think of but he also had his great mind.
He was interested in so many more things than just databases.
This is one of his famous challenges.
Give me 20 questions, 20 point questions that you want to ask of your day time.
参加者 2 1:28:22
It was the sky follow the strong digital sky surface.
This was one of the first massive data sets.
Well, there's a groundbreaking work on developing this guys here Jim's in that knowledge as a database expert.
What's transformative for the computation in the Sony data is actually a really funny story about that.
At some moment Jim goes to Philadelphia in Baltimore where the group Popius and he walks into the server Grou.
参加者 2 1:29:18
He knew that there was way too much mathematics.
This intuition builds from decades of experience that's given him the 6th sense of how systems should behave.
He tried to redesign their architecture and for improved dramatically
参加者 2 1:29:57
and like great technologist, I will describe him as teach deep around the main The through in understanding the skills you need to be successful in your job are a unique mix of personal skills, functional depth, industry knowledge.
A database developer who understands foot and performance or customer architectures can make better architectural choices because they see how they work shapes the overall system.
That breadth of knowledge gives you the perspective to improve what you build syst
参加者 2 1:30:53
there's breath they can dive deeper into a specific problems but they'll also understand how it fits into a larger system.
参加者 2 1:31:14
Contentative willingness to connect to multiple disciplines and ideas.
参加者 2 1:31:36
The experts in that field who understand that work fits into a larger system.
You must broaden your team.
Now if you look at sort of the renaissance development throughout this talk, I've called them different ways.
I think you need to be curious and keep learning thinking systems communicate with precision be an owner if you build it, you own it.
And finally, come not talk expand your knows you'll you have plenty of time next week to start putting all of this into practice before to night.
参加者 2 1:32:30
Join me at the Las Vegas 5th or Gance and will be grateful for some strezy games and of course live music.
Head like a vet and gate. It's a chance to celebrate with your teens and unwind after a week of serious learning and building.
Let us one more thing
参加者 2 1:32:59
I want you to leave you with this when you build something like, do you customers ever
参加者 2 1:33:56
in our world in the unseen systems that stay up for the night and clean deployments the robots that nobody notices most of what we built, nobody will ever see and the only reason why we do this well, it's our own professional prod in operational excellence.
That is what defines the best builders.
They do the function properly even when nobody is watching.
And when I look at the word that you delivers every day, I see that commitment everywhere.
参加者 2 1:34:44
So for that I am immensely proud of you.
Thank you for all that you do.
要約すると以下
Evolution of Software Development and Developer Qualities
全体の要約
The speech emphasizes the ongoing evolution of software development driven by curiosity, systems thinking, and ownership. It highlights how AI and rapid prototyping are revolutionizing the industry, enabling faster and higher-quality software creation. The speaker reflects on historical technological breakthroughs, the importance of learning from failure, and the need for engineers to adapt to change. It underscores the vital role of developers in addressing global challenges and advocates for operational excellence and professional integrity as essential for success.
主要トピック
- Evolution of software development
- Importance of curiosity and systems thinking
- Role of AI and rapid prototyping
- Historical technological advances
- Continuous learning and adaptation
- Global impact of engineering solutions
- Operational excellence and integrity
区間ごとの要約（長すぎるので折り畳み）
06:23 ~ 26:59 Navigating Development Tools and Voices - Navigating through ROM and Caesar forum discussed - Issues with program cards and compiler bugs raised - Advancements in programming languages highlighted - Evolution of development tools emphasized - Encouragement for younger engineers to share stories27:34 ~ 31:16
Evolving with Technology and AI
- Speaker stresses the need to evolve with AI
- Some skills may become obsolete, but adaptation is key
- Recalls education in the 60s and impact of compilers
- Late 1990s saw a shift to independent development teams
- Independence allowed for faster and more efficient work
31:54 ~ 35:24
Evolution of Development Practices and AI Impact
- Industry adopting new practices for distributed systems
- Developers in 2000s worked mainly on-premise
- Shift required new skills and adaptation
- AI assistants boost developer productivity
- Core work responsibility remains with developers
- Development landscape continues to evolve
- Current era seen as convergence of technologies
- Historical progress in one field aids others
35:54 ~ 38:35
Renaissance: Curiosity and Innovation Era
- Renaissance marked a shift from the Dark Ages
- Curiosity led to significant scientific advancements
- Inventions like the vanishing point transformed art
- Microscope and telescope were key inventions
- Printing press revolutionized information sharing
- Art and science were interconnected during this era
- Curiosity drove learning and experimentation
- Boundaries between fields were non-existent
39:25 ~ 42:06
Curiosity and Learning in Development
- Comparison of current era to the Renaissance
- Curiosity is essential for developers
- Continuous learning is necessary due to change
- Experimentation and willingness to fail are key
- Real learning comes from practical engagement
42:41 ~ 45:34
The Importance of Experiential Learning
- Experiential learning is crucial for understanding
- Balance of pressure affects performance levels
- Yerkes-Dodson law relates stress to learning
- Learning is enhanced through social interactions
- Travel experiences inform real-world technology use
- Community support initiatives can be profitable
46:18 ~ 49:39
Innovative Solutions for Health and Environment
- Efforts to clean oceans using AI and GPS
- Need to prevent new waste entering waterways
- Health Intelligence center processes vast healthcare data
- Data visualizes healthcare access and disease outbreaks
- Strategic placement of maternal health facilities
- Economic struggles of individuals highlighted
50:18 ~ 53:44
Innovative Tech Solutions for Global Challenges
- Introduction of ATM machines using ethanol
- Developers' role in solving global issues
- UN predicts 2 billion more people by 2050
- Technologists must ensure economic and healthcare solutions
- 265 heroes in technology across 58 countries
- Curiosity is essential for modern developers
- Systemic thinking is crucial for problem-solving
54:25 ~ 57:01
Understanding Complex Systems and Feedback
- Donella Meadows studied complex systems in the 70s
- Systems are interconnected entities producing behaviors
- Example: Wolves reintroduced in Yellowstone changed ecology
- Feedback loops can alter entire systems significantly
- Understanding requires viewing systems as wholes
- Changes in one part create new patterns in the system
- Reinforcing loops amplify change; balancing loops stabilize
- Recognizing patterns allows for effective interventions
57:39 ~ 1:01:50
Importance of Communication in Engineering
- Clear communication is vital for engineers
- Understanding systems requires a broader view
- Past example of Amazon architecture discussed
- Natural language can be ambiguous
- Need to reduce ambiguity in communication
1:02:19 ~ 1:04:20
Challenges in AI Communication and Spec Development
- User experienced communication issues with AI coding
- Prompts became longer, resembling software specifications
- Realized specs could improve AI interactions
- Sparked the idea for spec-driven development
- Initiated work on the hero IDE project
- Faced challenges in understanding user needs
1:04:55 ~ 1:06:33
The Role of Rapid Prototyping in Development
- Rapid prototyping aids in idea communication
- Historical example: invention of the mouse
- Prototypes provide better understanding than drawings
- AI enhances speed of graphic prototyping
- User feedback is crucial for development
1:07:03 ~ 1:08:39
Feature-Driven Prototyping with Ciro IDE
- Prototyping Ciro IDE for coding tasks
- Using CUR IDE to generate code efficiently
- Testing development steps with CVD techniques
- Kiira generates tests and application code
- Challenges in capturing software nuances
- Adopting traditional specs for detailed descriptions
- Voting tasks initiated by KERA for changes
- Long specs made it hard to locate changes
- Shifted to feature-driven specs for clarity
- Separated flow into requirements, designs, tasks
1:09:14 ~ 1:10:43
Enhancing Development with Curo IDE and AI
- Rapid prototypes provided critical user feedback
- CIRO ID and SPECLC enhance product development
- Ambiguous AI tasks lead to unclear outputs
- Spectrum development refines requirements precisely
- Curo IDE generates tasks before coding
- Example of production feature for notifications
- Addressed idle agent issue during coding work
1:11:07 ~ 1:12:27
Development of User Notification Feature
- Initiated feature for user notifications
- CUR generated stats to define agent actions
- Complex design led to new notification system
- Focused on external notification system development
- Built on Electron's native notification
- CUR spec workflow clarified necessary changes
- Integrated changes into agent code efficiently
- Project completed in half the expected time
1:12:58 ~ 1:14:40
Impact of ANI on Communication and Development
- ANI has transformed communication and iteration speed
- Rapid prototyping enhances user feedback collection
- Kiera IDE improves precision in natural language
- Effective communication reduces system errors
- Interview techniques focus on understanding customer needs
- Addressing customer queries requires deeper insights
1:15:16 ~ 1:17:30
Ownership and Quality in Software Development
- Customers must understand their problems and opportunities
- Communication between engineers and clients is essential
- Ownership of software quality is crucial
- AI can improve development speed and quality
- Risks arise from over-reliance on AI tools
- Developers are responsible for regulatory compliance
- Comprehension comes from active code creation
1:18:05 ~ 1:20:59
Challenges and Progress in AI Development
- Developers face challenges with AI-generated solutions
- Inappropriate or over-engineered proposals occur
- Progress in practices to improve code quality is noted
- Automated testing in CSCD pipelines is increasing
- Customer interactions are crucial for understanding needs
1:21:36 ~ 1:23:56
Mechanism for Consistent Outcomes in Engineering
- Introduction of a new mechanism for change
- Inspired by Toyota's Endum Ch principles
- Focus on defect prevention in production
- Mechanism converts intentions into outcomes
- Durability reviews enhance engineering practices
- Encourages consistent implementation of policies
- Creates a link between intent and execution
1:24:33 ~ 1:27:54
Integrating Human Judgment in AI Learning
- Integrate human judgment in software processes
- Encourage human-to-human coaching among engineers
- Diverse knowledge is essential for growth
- Leonardo da Vinci as a polymath example
- Mentor Jim Grave inspired broad curiosity
- Challenge to ask meaningful daily questions
1:28:22 ~ 1:31:36
Transformative Knowledge in Data Systems
- Importance of foundational knowledge in database systems
- Jim's expertise transforms data architecture
- Mix of personal skills and industry knowledge is crucial
- Understanding system components enhances design choices
- Encouragement for continuous learning and curiosity
- Call for team collaboration and project ownership
1:32:30 ~ 1:34:44
Celebration and Commitment at Las Vegas Event
- Join the Las Vegas event for games and music
- Celebrate and unwind after a busy week
- Unseen systems are vital to our success
- Best builders excel even when unnoticed
- Proud of the team's commitment and efforts
ありがとうございます。
Otter,ai ではなく、LINEの aiNoteサービスを利用しております。
現地で比較したので利点欠点などをあとでまとめます・・。