Skip to main content

Microsoft CEO Satya Nadella : 「知的コストのゼロ化」とソフト開発業界に起きる根本的な変革

· 35 min read

前置き​

最近のインタビューから。AI で要約させた。

音声対話(by AI)​

https://notebooklm.google.com/notebook/f52ab65a-c7cd-49d7-b078-9004bd6df059/audio

情報源 : 動画(20:00)​

Microsoft CEO Satya Nadella on the Future of AI

www.youtube.com/watch?v=w87UvmMcmW4

要約 by AI

インテリジェンスと技術の未来：主要テーマと考察

このブリーフィングドキュメントは、提供された英文ソース「インテリジェンスと技術の未来」からの抜粋を基に、インテリジェンスと技術の将来における主要なテーマ、重要なアイデア、および事実をレビューすることを目的としています。

1. AI時代における技術スタックの再構築​

最も顕著なテーマの一つは、AI時代における技術スタックの全面的な再構築の必要性です。AIエージェントの台頭は、インフラストラクチャ、データ、およびアプリケーションレイヤーを含む、技術スタックのすべての層に根本的な変化を要求しています。

  • インフラストラクチャの変革: AIワークロードは、以前のワークロードよりも多くのリソースを必要としますが、それは異なる単位のスケールでのことです。単にGPUやAIアクセラレータだけでなく、大量のストレージと従来のコンピューティング能力も必要とされます。これは、Azureの70のリージョンを「AIファクトリー」として適合させる必要性を生んでいます。
    • "Every layer of the tech stack has to be reimagined." "So take even the infrastructure layer, right? Obviously, we are very, very proud of the fact that we have 70 regions of Azure all over the world. And then you say, wow, we now need to retrofit them or fit them to become AI factories, right? That's kind of what you need to do." "Interestingly enough, it turns out what we built over the last 15 years may be more relevant now because agents need more of it than any workload previously. But it's at a different unit of scale."
  • データ層の変化: 従来のデータベースはデータの構造化に焦点を当てていましたが、これからはデータに「インテリジェンス層」をもたらすことが重要になります。SQLクエリにLLMレスポンスを組み込むデモが示唆するように、データへの推論エンジンの組み込みが求められています。
    • "But now you can bring the intelligence layer to the data, right? A reasoning engine straight in." "One of the coolest demos we showed was this Postgres, which is so modular that you can now mix and match in your SQL query an LLM response, right?"
  • アプリケーション層の「エージェントへの崩壊」: ソフトウェア、特にアプリケーションレイヤーは、エージェントに「崩壊」していくという考えが提示されています。これは、従来のSaaSアプリケーションが単なる「システムの記録」や「システムのエンゲージメント」であることから脱却し、エージェントが複数のバックエンドシステム（SaaSアプリケーションを含む）をオーケストレーションする新しいフレームワークに参加する必要があることを意味します。
    • "You had previously said the software, the application layer is going to collapse down into agents." "So I think we have to all be open to participating in what is this new orchestration layer in the agentic web that will be, essentially, will have multiple backends, right? Your SaaS application will be one backend."

2. エージェント中心の新しいユーザー体験とMicrosoft 365の進化​

AIエージェントの統合は、エンドユーザーの製品体験を根本的に変えつつあります。Microsoft 365を例に、新しいユーザー体験のモードが説明されています。

  • AIのための新しいUI: チャット、検索、ノートブック、そしてタスクを委任できる「エージェント」（研究者やアナリストのような）を含む、AIのための新しいユーザーインターフェースが登場しています。Copilot Studioにより、ユーザー自身がエージェントを構築することも可能です。
    • "The first is the brand new mode, which is I have this new UI for AI, which literally is this new scaffolding, which has chat, which has search, which has notebooks, which is a place where I collect all these heterogeneous collections of data and do things like podcasts and audio overviews and all of that. I have agents, right? Like these researchers and analysts."
  • マルチプレイヤーモードでのAI: Teamsのようなプラットフォームは、エージェントを「マルチプレイヤーモード」に持ち込みます。チャネルやミーティング内でエージェントが利用可能になり、チームでの協力を強化します。
    • "Teams takes all of that into multiplayer mode, right? All of those agents are available to me in my channel, in my meeting, right? So Teams becomes the scaffolding in which all of the AI now is working with me in multiplayer mode."
  • IDEとしてのOfficeキャンバス: 各Officeアプリケーションは、Copilotの統合により、チャット機能を備えた統合開発環境（IDE）のように機能します。例えば、Excelでの作業中にデータサイエンティストが隣にいるような体験や、文書作成中に研究者がいるような体験が実現されます。
    • "The idea is we have turned every office canvas into an IDE with chat, essentially, if you think about it that way."

3. 企業におけるエージェントの知的財産と管理​

従業員が個人的なエージェントを業務に持ち込む可能性について議論されており、企業の知的財産（IP）とセキュリティに関する重要な考慮事項が強調されています。

  • エージェントのIP所有権: 従業員の仕事の成果が企業の財産であるのと同様に、エージェントも企業の知的財産になると考えられています。
    • "Actually, that's correct. In fact, what you said is sort of our view, right? If I look at even today's announcements, because what is the intellectual property of a company, right? Even the work product of any one of us at work is the company's property. And so that I think is going to be the case even with agents."
  • ID管理とセキュリティ: エージェントが企業データにアクセスするため、人々とITインフラストラクチャのために行われてきたのと同じID管理、データ保護、セキュリティ対策（アクセス管理、Purview、Defenderなど）がエージェントとそのITインフラストラクチャにも適用される必要があります。
    • "So absolutely, thinking of all of the things we've done from identity management and security for people and their IT infrastructure is going to be done for agents and their IT infrastructure."
  • 個人と企業の領域の分離: 個人的なエージェントを業務に持ち込む場合、データ漏洩を防ぐために、これらの世界を分離することが不可欠です。これは、Microsoft AccountとEntra（旧Azure Active Directory）を用いたCopilotとMicrosoft 365 Copilotのように、異なるIDと状態を分離する既存の仕組みと同様の考え方に基づいています。
    • "The system where you bring your own personal agents, right, has to be done in such a way that these two worlds don't have the data leakage, right?" "For both privacy reasons and for also intellectual property reasons, right? Both of those are helpful. I think something like that. That's why even we believe in Entra and Microsoft Account, right?"

4. 知的コストのゼロ化がもたらす影響と社会への貢献​

知的コストがゼロに近づくことの可能性と、それが社会にもたらす潜在的な影響について楽観的な見方が示されています。

  • 生産性と経済成長の促進: 知的コストの低下は、生産性と経済成長を促進するための「助け」として期待されています。特に、インフレの抑制や経済成長の向上といった課題に対処する上で重要です。
    • "What are you most excited about when the cost of intelligence approaches zero?" "Yeah, I mean, to me, ultimately, right, when I look out there in the world, do we need more abundance of something like technology, like intelligence that can then ultimately drive productivity and economic growth? Absolutely, right." "You know, right now, to tame either inflation or improve economic growth, we need some help. Like, so this is, you know, high time."
  • 高付加価値分野への応用: スタンフォード大学医学部での腫瘍カンファレンスでのAI活用事例のように、ヘルスケアなど高付加価値かつ社会的に重要な分野でのAIの応用が進むことが期待されています。これは、患者ケアの改善、アウトカムの向上、コスト削減に繋がる可能性があります。
    • "So if you sort of take that thing and take sort of the example we even shared in our developer conferences to what Stanford Medicine was able to do for something real high stakes, right, like, you know, tumor board meetings and oncology and cancer care... That is, to me, what needs to happen. Healthcare is 20% of our GDP, close to a lot of the expenses in this workflow. And so if, you know, every provider out there can start improving their patient care and outcomes and reducing costs using AI, that's going to have a profound impact in our societies."
  • 広範な分野への波及: AIはヘルスケアだけでなく、マテリアルサイエンス、幅広い知識労働、中小企業の生産性向上など、様々な分野で価値を創出することが必要とされています。これは、エネルギー使用量の増加に対する「社会的な許可」を得るためにも重要です。
    • "It has to be in health care, in material science, in broad knowledge work in a small business getting productivity, because that's what will give us the social permission to continue to use the scarce resource called energy."

5. AIのエネルギー使用量と持続可能性​

AIのエネルギー使用量に対する懸念、特に若い世代からの懸念が認識されており、Microsoftのアプローチが説明されています。

  • 懸念の認識と応答: AIのエネルギー使用が地球に著しく悪影響を与えるという懸念があることを認めつつ、若い世代がこの問題に深く関心を持っていることを賞賛しています。
    • "anecdotally, I've heard from some of the younger generation that they're either avoiding AI altogether or maybe just using it lightly and specifically because the energy usage they're thinking is going to pretty significantly negatively affect our planet." "Yeah, first of all, you know, the fact that the younger generation cares about this deeply is so, so inspiring, quite frankly..."
  • 持続可能なアプローチ: 技術創造の目的が社会の課題解決（ヘルスケア、教育など）にあることを強調しつつ、それを「持続可能な方法」で行う必要性を訴えています。これは、「持続可能な豊かさ（sustainable abundance）」という概念で表現されています。
    • "Then the second part is also important, which is we've got to do it in a sustainable way, right? It's sustainable abundance."
  • 効率性の指標と再生可能エネルギー: 効率性の指標として「Tokens per dollar per watt」を挙げ、エネルギーを最も効率的に使用して最大の豊かさを生み出すことの重要性を指摘しています。また、Microsoftが再生可能エネルギーの最大の購入者の一つであり、新しいプロジェクトへの補助金の提供者であることにも言及し、持続可能性へのコミットメントを示しています。
    • "one of the equations I go back to is it's tokens per dollar per watt, right?" "We are some of the biggest buyers of alternative energy. In fact, you could say the greatest subsidy out there for new projects is from people like us."

6. オペレーティングシステムの未来：決定論的コードとエージェントの融合​

決定論的な伝統的なコードと非決定論的なAIエージェントの境界線が曖昧になる可能性について、オペレーティングシステムの将来を念頭に置いた議論が行われています。

  • ゲームにおけるAIの活用例: ゲームのフレーム生成にディフュージョンモデルを使用したり、Xboxコントローラーの操作で次のシーンを生成したりする例が示され、全てが生成されるシステムの可能性が示唆されています。
    • "I saw this really cool demo a few months ago where they recreated the game Doom using a diffusion model. Every single frame was predicted." "We had a similar one, like this Muse model that we built was a world action model that we had built, but we trained it on gaming data... Everything is generated, so to speak."
  • システムの理解と制約: 完全に非決定論的なシステムではなく、確率的なシステムであっても「検査可能な」決定論的な方法で機能する必要があると述べられています。これは、「知性の物理学」を理解し、複雑なシステムを結合する際にそれを「制約（bound）」することの重要性を示唆しています。
    • "Yes, it's a stochastic system, but this stochastic system does need to work in deterministic ways that we can at least, like, you know, inspect." "Which is we have to somehow get back to a place where when we stitch these complex systems, there is some way for us to understand the physics of these complex systems, and then bound them, right?"
  • エージェント環境の管理: コーディングエージェントの例を挙げ、エージェント環境を仮想マシンとして扱い、インターネットアクセスやツールへのアクセスなどを制御し、完全な監査ログを持つことの重要性を説明しています。これは、従来の命令型コードとエージェントを組み合わせて、そのインタラクションを監視する方法を示唆しています。
    • "Interestingly enough, the coding agent has an environment. Underneath GitHub Actions, you're really putting a virtual machine, and then you're really setting the boundaries of that virtual machine, right?"

結論として、 この抜粋は、AI、特に強力なエージェントの台頭が、技術スタックの根本的な再構築、ユーザー体験の変革、企業のIT管理方法の変化を推進していることを示しています。同時に、知的コストのゼロ化がもたらす生産性の向上と社会への貢献への期待、そしてエネルギー使用量と持続可能性への配慮の重要性も強調されています。技術の進化はまだ「中盤」であり、決定論的なシステムと非決定論的なシステムの融合がどのように進んでいくのか、今後の展開が注目されます。

文字起こし

The energy usage, they're thinking, is going to pretty significantly affect our planet. Agents need more of it than any workload previously, but it's at a different unit of scale. What are you most excited about when the cost of intelligence approaches zero? You know, right now, to tame either inflation or improve economic growth, we need some help. Do you ever see a future in which the operating system is similar, where there's maybe no very little underlying traditional code? (0:00:30)

Every layer of the tech stack has to be reimagined. You had previously said the software, the application layer, is going to collapse down into agents. And what does that mean for vertical SaaS companies? All right, Satya, thanks for chatting with me. Congratulations on everything you announced at Build. Yeah, I had a few questions. You've overseen very successful major transitions at Microsoft Cloud, the adoption of open source, and now we're kind of in this next one. (0:01:02)

With the rise of these incredibly powerful AI agents, how are you thinking about investing in that while also maintaining the current product suite that you have and the major changes that are coming to it? Yeah, first of all, thanks so much for being at our developer conference. You know, the way I think about it is, first, you have to embrace what's new, right? (0:01:26)

And that's where I think even now that we're even two to three years in, depending on how you count into this AI era and this agentic web era, the patterns of what it means to build agents, build apps is becoming clearer, right? So you have to really go look at this tech stack that you have, that you built maybe for a previous workload now needs to be really thought of from first principles for the new workload, right? (0:01:56)

So take even the infrastructure layer, right? Obviously, we are very, very proud of the fact that we have 70 regions of Azure all over the world. And then you say, wow, we now need to retrofit them or fit them to become AI factories, right? That's kind of what you need to do. Turns out that if you take even an app like ChatGPT or Copilot, it definitely needs a lot of GPUs or AI accelerators, but it also needs everything else. (0:02:23)

It needs, in fact, tons of storage, both in during training and inference. It needs a ton of compute, regular compute, not AI accelerated compute, in order to be able to, for example, have the environments for agents. So interestingly enough, it turns out what we built over the last 15 years may be more relevant now because agents need more of it than any workload previously. (0:02:47)

But it's at a different unit of scale. So that's kind of what we have to do at the infra layer. Same thing with data, right? Take data and you say, well, you know, data has been always about... take a database. You said that's where you schematize people, places, things. (0:03:01)

But now you can bring the intelligence layer to the data, right? A reasoning engine straight in. One of the coolest demos we showed was this Postgres, which is so modular that you can now mix and match in your SQL query an LLM response, right? I mean, think about the query plan you can generate. So I feel that every layer of the tech stack has to be reimagined, but it also means we can take some of the best work we've done over the last 15 years and have it compound for our developers so that they can get the benefits of it. (0:03:37)

That's how we're thinking about it, which is how do we make sure we think about every layer of the tech stack from a first principles perspective for the new AI workloads that are being built and then really stitch together so that it meets the real world needs of customers. So the end users, especially for products that people are so familiar with, Office 365, these products, I assume, are going to be changing so rapidly. (0:04:01)

So what is that acceleration in the change in the product? Yeah, it's a fascinating thing, right? If I look at Office, there are, I'd say, three modes of Microsoft 365. The first is the brand new mode, which is I have this new UI for AI, which literally is this new scaffolding, which has chat, which has search, which has notebooks, which is a place where I collect all these heterogeneous collections of data and do things like podcasts and audio overviews and all of that. (0:04:33)

I have agents, right? Like these researchers and analysts. So these are things that I can delegate tasks to and so on. It's so exciting to have all of that. And I even have the co-pilot studio. So in other words, I can build agents, right? So that's the new thing, which is, you know, like I now have a UI for AI and agents. Here's the other interesting thing. (0:04:54)

Teams takes all of that into multiplayer mode, right? All of those agents are available to me in my channel, in my meeting, right? So Teams becomes the scaffolding in which all of the AI now is working with me in multiplayer mode. And the third mode is my heads down, just like how in GitHub co-pilot with VS Code, I'm heads down coding away, but I have agents that I can use. (0:05:18)

I'm heads down on an Excel spreadsheet and I have, you know, my co-pilot chat right there, right? That is like having a data scientist next to me while I'm analyzing a spreadsheet. While I'm research, writing a document, I have a researcher right there. The idea is we have turned every office canvas into an IDE with chat, essentially, if you think about it that way. (0:05:40)

So in some sense, I feel like the value of even the M365 system now has gotten to be more compounded because of intelligence getting built into all of these layers. (0:05:53)

So I want to continue on that note a little bit. You had previously said the software, the application layer is going to collapse down into agents. And I made a video about it, called it SaaS is dead. It garnered a lot of attention. That was really cool thoughts about it. But I want to hear, you know, the assertion is there's going to be the agent layer. (0:06:15)

And then under that, there's going to be the grounded database that the agents can read and write from. And what does that mean for vertical SaaS companies? How do they prepare themselves for this future? Yeah, like all of us. I think the way to go about it, I mean, even the demo we showed today, right? Literally, there was Dynamics 365 with essentially an MCP server that was used by Copilot Studio to orchestrate a multi-agent application that spanned CRM and many other systems of record and then completed, essentially, the orchestration of a complex business process, right? (0:06:54)

That seems to be here and now, right? I mean, it's kind of clear as day that when you go about thinking about business process and business applications, you have to compose yourself into something like that. And so, yeah, that will mean pretty radical change if all you thought was, hey, I'm the system of record or system of engagement or what have you. And it's just about workflows on top of my data. (0:07:19)

And that's the scope. That's just not going to persist. So I think we have to all be open to participating in what is this new orchestration layer in the agentic web that will be, essentially, will have multiple backends, right? Your SaaS application will be one backend. You better support something like MCP in order to be able to participate in the agentic web. And then maybe even something like NLWeb could even reduce the friction of all these connectors, right? (0:07:48)

Because if you think about enterprises, you have a significant sort of friction in how connectors work. Something like NLWeb could be massive change even inside the enterprise. And so I feel like, yes, I think SaaS applications we built may have to radically change in order to adapt themselves for this future. Yeah. And so these, let's just say these SaaS companies, do you think they're going to be curators of this ground truth data for their customers? (0:08:15)

And then the agents will be provided by platform companies like Microsoft? Yeah. I mean, I sort of think about like, it's unclear to me exactly how it all shakes out because in some sense, all of us overstate the importance of one thing that we have today. But the reality is in these platform shifts, the value will anyway get created somewhere else, right? So this is where, you know, like at the end of the day, what's the job to be done? (0:08:38)

The job to be done is to complete a business process. It's not about any one system of record and its curation, or it's not about any one agent or one workflow. (0:08:48)

It's about the entirety of it, right? So that to me is where the water is flowing. And the question is, how do you flow with it versus thinking that somehow I have this moat and I'm going to hold on to it, or I'm going to build some sort of facade around it, which has an agent head or what have you, but it's not the comprehensive view of what the customer in this case needs to get done. (0:09:13)

Yeah, I really like that you said, you know, different types of agents are going to be talking together, different databases, it doesn't really matter. It's just an abstraction layer. So that all sounds really exciting. Another thing you mentioned, you had said in an interview that when you hire somebody and an employee, you're hiring them in the future and their basket of agents, which is fascinating to me. (0:09:36)

But I want to get some clarity on that, because it seems like most likely the company, the employer is going to want to own that IP, the agents, just like they would own traditional IP. So maybe you can add some clarity. Actually, that's correct. In fact, what you said is sort of our view, right? If I look at even today's announcements, because what is the intellectual property of a company, right? (0:09:58)

Even the work product of any one of us at work is the company's property. And so that I think is going to be the case even with agents. And that's one of the fundamental reasons why we extended the rails, right? Now agents have an entire ID. You can manage conditional access for these agents, just like you do with people inside of an organization. (0:10:23)

Purview, another super important thing. If agents are going to have access to data, it needs to be subject to the same data protection and data rights. We are going to... and security, by the way. You want to manage the agent environment like an endpoint. So that's why Defender is going to make sure that there is no credential theft or what have you. So absolutely, thinking of all of the things we've done from identity management and security for people and their IT infrastructure is going to be done for agents and their IT infrastructure. Yeah, you know, that makes a lot of sense. (0:10:57)

And I also suspect a lot of people are going to build their own personal agents for their personal life. And maybe, do you see a future in which they are also bringing those personal agents into work? That's a great question. I mean, the system where you bring your own personal agents, right, has to be done in such a way that these two worlds don't have the data leakage, right? (0:11:19)

So that's the issue, right? Which is, you know, let's even take the simple thing, which is, hey, my email and my corporate email. Today, they're two segregated things. They're two identities. And we know how to separate state out. (0:11:31)

For both privacy reasons and for also intellectual property reasons, right? Both of those are helpful. I think something like that. That's why even we believe in Entra and Microsoft Account, right? So that's why we have Copilot and Microsoft 365 Copilot. Conflating the two, even in terms of the user model, it can be very confusing. The reason why pin edge with two profiles is helpful, because I do it with Microsoft Account as a user, as an individual. (0:12:03)

And I do this with Entra when I'm working at Microsoft. And it's a helpful demarcation to keep this, you know, the mental model simple. Otherwise, I think by conflating it all, I think you can really get very tangled. I think that makes a lot of sense. So I want to ask you maybe a question about your vision. So the cost of intelligence does seem to be dropping rapidly, hopefully approaching zero. (0:12:26)

I think it's going to be such a fascinating world in the future. What do you think, what use cases do you think are going to open up? What are you most excited about when the cost of intelligence approaches zero? Yeah, I mean, to me, ultimately, right, when I look out there in the world, do we need more abundance of something like technology, like intelligence that can then ultimately drive productivity and economic growth? (0:12:53)

Absolutely, right. I look around and say, man, you know, right now to tame either inflation or improve economic growth, we need some help. Like, so this is, you know, high time. So if you sort of take that thing and take sort of the example we even shared in our developer conferences to what Stanford Medicine was able to do for something real high stakes, right, like, you know, tumor board meetings and oncology and cancer care, they were able to take all this technology and apply it in a real way where it was a multi-agent framework they used in Foundry to orchestrate, you know, pathology, clinical trials, PubMed data, ultimately to have a better tumor board meeting and then have the data and the information from that going to PowerPoint into a teaching class or into teams for people to collaborate. (0:13:43)

That is, to me, what needs to happen. Healthcare is 20% of our GDP, close to a lot of the expenses in this workflow. And so if, you know, every provider out there can start improving their patient care and outcomes and reducing costs using AI, that's going to have a profound impact in our societies. And so that's what I'm really looking forward to. Yeah, I think the healthcare use case is, I'm extremely excited about hyper-personalized healthcare. I already used ChaiGPT co-pilot to answer questions about my own personal health. (0:14:18)

It is very, very exciting. And also some of the research that you showed off, the immersion cooling, right? That was discovered. Yeah, it's so cool. Material science, there's just so much there. So anecdotally, I've heard from some of the younger generation that they're either avoiding AI altogether or maybe just using it lightly and specifically because the energy usage they're thinking is going to pretty significantly negatively affect our planet. (0:14:49)

And so how do you, how does Microsoft think about that? (0:14:53)

What words of confidence would you give to them? Yeah, first of all, you know, the fact that the younger generation cares about this deeply is so, so inspiring, quite frankly, because at some level, it is sort of the right push for all of us to be able to say, OK, whatever we are creating is fundamentally helping with these outcomes that matter to us in society, right? (0:15:16)

Whether it's in health care, whether it's in education, whether it's access to financial services, whatever the domain is, ultimately, it's economic growth, it's economic prosperity and abundance. So let's take that as the first thing, which is we're not doing this just for some tech accomplishment, but we're doing it to solve the challenges of people and planet. Then the second part is also important, which is we've got to do it in a sustainable way, right? (0:15:40)

It's sustainable abundance. And then if that is the case, then one of the equations I go back to is it's tokens per dollar per watt, right? The fact that we are able to use software, right, as the most malleable resource to use energy in the most efficient way to generate the most amount of abundance, which in turn then improves health and education and other outcomes is a good thing. (0:16:09)

And we just have to stay on it, right? The reality is all of tech is what 2%, 3% of sort of the grid power today or total power consumption. So it's minor, but yes, it'll double. And so if it needs to double, it needs to have the social permission to double. It needs to create a lot more value in the real world. In fact, that's one of the reasons why I feel like we as a tech industry need to be anchored on not just one product of ours that everybody's using for doing fun things. (0:16:41)

It has to be in health care, in material science, in broad knowledge work in a small business getting productivity, because that's what will give us the social permission to continue to use the scarce resource called energy. Do that in a sustainable way. We are some of the biggest buyers of alternative energy. In fact, you could say the greatest subsidy out there for new projects is from people like us. (0:17:07)

We really want to make sure we continue to push on it, but ultimately deliver tokens per dollar per watt such that it creates economic prosperity. I'm glad you said that. I'm going to tell the folks who are very nervous about the environmental impact what you said, and I'll show them this video. So thank you. There's definitely a major shift in computing architecture happening right now. (0:17:29)

The line between deterministic and non-deterministic parts of the architecture are kind of getting blurred. I saw this really cool demo a few months ago where they recreated the game Doom using a diffusion model. Every single frame was predicted. Do you ever see a future in which the operating system is similar, where there's maybe no or very little underlying traditional code? (0:17:53)

That's a good one. We had a similar one, like this Muse model that we built was a world action model that we had built, but we trained it on gaming data. And it starts to do. Essentially, you can have an Xbox controller as actions that are then used to generate the next scene, which is sort of pretty... It's kind of like robotics, you can think of it like that, and gaming is like that. (0:18:15)

And so, yes, everything is generated, so to speak. The operating system, you know, to me, this idea that, you know, sometimes I think we even overstate the determinism of what we call deterministic systems, right? Because after all, if you take a software program, you can't prove it, right? I mean, that's sort of one of the fundamental challenges of computer science, that it's... And so, therefore, I think we... Yes, it's a stochastic system, but this stochastic system does need to work in deterministic ways that we can at least, like, you know, inspect. (0:18:55)

Quite frankly, what Elon said in the keynote when I interviewed him, he said, hey, we got to understand the physics of intelligence. It's actually a good way to think about it, right? Which is we have to somehow get back to a place where when we stitch these complex systems, there is some way for us to understand the physics of these complex systems, and then bound them, right? (0:19:16)

Sandbox them, and what have you, right? I think that that's what we will have to do even in an operating system. But when I look back at it, you take the coding agent that we just launched. Interestingly enough, the coding agent has an environment. Underneath GitHub Actions, you're really putting a virtual machine, and then you're really setting the boundaries of that virtual machine, right? (0:19:34)

Which is, hey, does it have internet access or not? You have to control it. If it's going to have access to tools with MCP servers, you have to control it. And then there's a full audit log of all of that. And so I think that that's how we will learn how to essentially mix a deterministic system, so to speak, which is a software system that we built with a lot of imperative code, and this agent, to couple together, and then have the interaction itself be something that we can monitor. (0:20:04)

Yeah, so it's really cool, because you said in the keynote, we're kind of beginning the middle innings of this transition. So I really think it's a really interesting time to see where that blend of different types of software go. Thank you so much for meeting with me. I appreciate it. Thank you so much. Such a pleasure. Thanks for coming and spending the time. (0:20:24)

And I look forward to being in touch.

(0:20:26)

(2025-05-28)