ヤン・ルカンが語るLLMの先にあるもの

You're one of the godfathers of AI. あなたはAIの父の一人ですね。 What's your kind of view of the path of progress here? この分野での進歩の道筋について、どのようにお考えですか？ 5 years complete world domination. 5年で完全な世界制覇です。 The best way to get breakthrough research is you hire the best people. 画期的な研究を生み出す最善の方法は、最高の人材を採用することです。 You get the out of the way. あとは邪魔をしないことです。 Pardon my French. 失礼な言い方をお許しください。 You share the touring award with two others. チューリング賞を他の二人と共同受賞されましたね。 When did your views start diverging? お二人との意見の相違はいつ頃から始まったのですか？ In 2023. 2023年です。 How do you know it was time to leave Meta? Meta を離れる時期だと、どうやって判断されたのですか？ It sounds like you were thinking through some of these things over a period of time. ある程度の期間をかけてこうしたことを考え続けてこられたようですね。 Here's a big misconception about my role, my relation to Alex, and how AI was run at Meta. 私の役割、Alex との関係、そして Meta における AI の運営方法について、大きな誤解があります。 What's like one thing you've changed your mind on in the last year? この一年で考えが変わったことを一つ挙げるとしたら？ I mean, the whole idea of uh そうですね、その概念全体について言うと Yan Lun is one of the godfathers of AI. Yann LeCun はAIの父の一人です。 He's an absolute legend in the field. この分野における紛れもない伝説です。 Uh someone I've admired for a long time. 私が長い間尊敬してきた方です。 And so it was such a treat to get him on unsupervised learning. Unsupervised Learning にお招きできて本当に光栄です。 Uh he's been a noted skeptic of of LMS in many ways. LLM の懐疑論者として知られています。 And so we dug into what LM can do, what they can't do, uh some of the limitations he sees, and why he ultimately decided to pursue a different architecture. LLM にできること、できないこと、その限界について深く掘り下げました。 Uh and we also talked about his time at Meta. Meta での時代についても話を聞きました。 um you know the things he's proud of in setting up fair how the last few years proceeded and what ultimately led him to uh spin out and start his own company uh AMI um I think it's just fascinating to get Yan's thoughts on everything happening in the AI ecosystem today this tension between basic research and then pushing LLM forward and how that's happening in in a bunch of organizations today as well as his thoughts on just where the the whole space is headed uh he's just an absolute giant in the field and when I started this podcast I hope we get guests like him so it is just such a treat I think folks will really enjoy hearing the conversation we had. FAIR 設立で誇りに思うこと、この数年間の経緯、最終的に独立して AMI を立ち上げるに至った経緯を語ってくれました。LLM の前進と基礎研究の緊張関係や、AI エコシステム全体の行方についての彼の考えを聞けるのは本当に貴重です。 Without further ado, here's Yan. それでは、Yann をお迎えしましょう。 Yan, this is such a pleasure. Yann、お会いできて光栄です。 You're one of the godfathers of AI. あなたはAIの父の一人ですね。 I feel like when I started doing this podcast years ago, I was really hoping we might one day get someone like you on. このポッドキャストを始めた数年前から、いつかお招きしたいと思っていました。 You know, I don't like that term because I live in New Jersey. その呼び名は好きじゃないんです、ニュージャージーに住んでいるので。 When you're a godfather in New Jersey, ニュージャージーでゴッドファーザーといえば、 it doesn't mean the same thing. 同じ意味にはなりませんからね。 Very fair. それはもっともです。 Very fair. それはもっともです。 You know, obviously, you know, your bet on on neural nets when everyone doubted them is legendary. 誰もが疑っていたときにニューラルネットに賭けたことは伝説的ですよね。 And I feel like today you're making a similar bet in many ways against LLMs and the kind of predominant generative architectures that that so many believe in. そして今度はある意味で LLM に逆らうという同じような賭けをされているように見えます。 Uh you've recently started a new company uh behind this theme. 最近このテーマを掲げた新会社を立ち上げられましたね。 And so you know our goal today in the conversation is to leave our listeners with a lot more information about AME, what you're doing there, some of your work at Tapestry. 今日の対話の目標は、リスナーの皆さんに明確なビジョンを持って帰っていただくことです。 Um you know, why you think the rest of the field is is is pointed in the wrong direction around some of these generative models and then also just get your reflections on the way the field's unfolded your time at Meta and all that. なぜ残りの業界が生成モデルで間違った方向を向いていると思うのか、Meta での時代の振り返りなど聞かせてください。 So, you know, modest goals for uh for for for a single podcast episode. 一回のポッドキャストとしてはなかなか大きなテーマですね。 I figured it'd be great to start with the meat um because the company feels like the clearest statement of your technical thesis going forward. 会社の話から始めるのがいいと思いました。会社そのものが論文のように感じられたので。 And so, you recently launched the company. 最近会社を立ち上げられましたね。 It's focused on world models uh and scaling the Jeter architecture, which you obviously pioneered uh over at Meta. AMI は世界モデルと、あなたが Meta でパイオニアとして取り組んできた JEPA アーキテクチャのスケーリングに注力しています。 And so, I'm wondering if you could talk a little bit about the origins of that architecture and the extent to which you drew inspiration from the human brain and the way that works. このアーキテクチャの誕生背景について少し話していただけますか？ So first of all, I want to say there's nothing wrong with LLMs in the sense of LLM, you know, are the basis for a lot of very useful AI products that all of us use, including me. まず言っておきたいのですが、LLM 自体に問題があるわけではありません。 Uh they're great, okay, for what they do. その用途においては素晴らしいものです。 They're just not a path towards human level or human like intelligence or even animalike intelligence. ただ、人間レベルの知能や動物レベルの知能への道筋にはなりえないのです。 Uh so that's my claim, okay? それが私の主張です。 I'm not saying are useless, right? 役に立たないと言っているわけではありません。 I'm I'm just saying they're not a path towards you. ただ、そちらへの道筋にはなりえないと言っているだけです。 I mean, つまり、 you helped build some of the first major open source ones, 最初の主要なオープンソースモデルの一部を構築されましたよね。 right? そうですね。 Absolutely. もちろんです。 So, what is uh AME? では、AMI とは何ですか？ So, ME really stands for advanced machine intelligence and the the the kind of subtitle the moto if you want is uh AI for the real world. AMI は Advanced Machine Intelligence の略で、その種のシステムが必要とするのは So basically a lot of you know AI techniques that people know about today are good for language manipulation either human language or computer code or mathematics or or legal ease which barely qualifies as human language. 今日人々が知っている多くの AI 技術は言語を前提として設計されています。 Unfortunately a lot of human language used for it 残念ながら人間の言語がその多くに使われています。 right sadly you know language is very special in a way and it's particularly well suited for the type of uh you know architectures that have been so successful uh recently the the you know large language models GPT style architectures but what about the real world what about like understanding the physical world turns out reality is way more complicated than language uh because 言語はある意味で特別で、予測モデルの学習に非常に適しています。 It's highdimensional. 高次元です。 It's continuous. 連続的です。 It's noisy. ノイズが多いです。 It's messy. 乱雑です。 And uh training a system to understand the real world is much much harder. 現実世界を理解するシステムを学習させるのははるかに難しいのです。 So that's really what we're after. それが本当の目標です。 That's what I've been after for most of my career. それが私のキャリアのほとんどで追い求めてきたことです。 And really kind of, you know, working on in an accelerated fashion over the last five, six years or so and making significant progress over the last two years. この数年で加速した形で取り組んできたことです。 And so it made sense to really do a startup around it and sort of go to into high gear, you know, in pushing that. それを中心にスタートアップを立ち上げ、より集中して進めることが理にかなっていました。 and it became clear, you know, by the end of last year that Meta was really not the right place for that. 昨年末には Meta がその方向に本気で向かっていないことが明らかになりました。 So, which is why I left and started Emmy Labs. だから私は Meta を離れ、AMI Labs を立ち上げたのです。 I think it's an interesting like, you know, trend that we're seeing across the board, right, where it feels like um there you're there's there's many folks spinning out of, you know, either some of the large companies or research labs, you know, that have a a particular direction of research they're excited about. 業界全体で見られる興味深いトレンドだと思います。 And you you'd have such an interesting vantage point of this from your time at fair. FAIR での経験があるから、独特な視点を持っていらっしゃいますよね。 This uh almost tension that exists between, you know, go pursue as many different research directions as possible in these companies versus hey, something's really working. できるだけ多くの方向性を追求するという研究の姿勢と、製品化に集中するという緊張関係がありますね。 This is the thing that we're going to sell for the next 61 12 months like go focus on that. 次の6〜12ヶ月はこれを売り込んでいこう、という絞り込みの必要性とのせめぎ合いです。 You know, I'm curious your your thoughts on that and and what you've kind of seen in the industry at large. その点についてのお考えや、これまでの経験を聞かせていただけますか？ Well, it's a strange uh trade-off. 難しいトレードオフですね。 There's really two modes of R&D, right? R&D には大きく二つのモードがあります。 There's a lot of exploratory research, a lot of d research directions, right? 探索的な研究と、多くの研究方向性があります。 And sometimes something kind of seems to work and you you need to push it further and it's not research anymore. 何かうまくいきそうなことが見えてきたとき、さらに進めていく必要があります。 I mean the people working on it are researchers or they're called researchers at least in the press but uh but really it's becoming more engineering and pushing for for products, right? 取り組んでいる人々は研究者、あるいは大企業では研究者と呼ばれています。 So that happened a number of times at Meta because of things that was started at fair. Meta でも、FAIR から始まったことが発端でそういうことが何度かありました。 Such a thing happened in you know early 2023 essentially uh when you know Lama which was developed at fair lama one um was very promising and uh meta created a whole organization geni to turn it into something real and a series of products uh and produced you know lama 2, lama 3, lama 4 which was a bit of disappointment uh and because you know Mark Zuckerber was disappointed by it he kind rebooted the entire organization, reorganized it and hired new people etc. 2023年初頭、LLaMA が公開された頃にそういうことが起きました。 But what also happened uh over the last year is that uh basically the company meta realized that um they had fallen behind a little bit and so that kind of refocused the the strategy on trying to catch up with the industry. 昨年に起きたもう一つのことは、会社が基本的にその事業の取り組み方を変えたことです。 And the sad side effect of it is that a lot of the exploratory research was basically not given high priority anymore. その悲しい副作用として、探索的な研究の多くが基本的に打ち切られてしまいました。 I mean it didn't concern the stuff I was working on. 私が取り組んでいたことには関係なかったのですが。 all the JA and role models because you know Mark himself and and Dub Bosworth the CTO and a bunch of other people in the company were really interested in that project and really believed in the long-term impact but the rest of the company was just you know totally entirely focused on LLM and made it clear to me that Ma was really not the the right place to push for that project anymore and then we started to had good results and so it was clear that you know we had to kind of make that transition between research and actually kind of uh developing the technology, scaling it up and building products out of it. JEPA と世界モデルについては Mark 自身と Andrew Bosworth、そして会社の多くの人が非常に乗り気で長期的な影響を信じていましたが、残りの会社は完全に LLM に集中していて、Meta がそのプロジェクトを推進する場所として適切でないことが明確になりました。 And we realized also that most of the applications were probably for things that Meta was not particularly interested in. 多くのアプリケーションは、おそらく Meta が関心を持っていないことのためのものだとわかりました。 A lot of applications of the kind of stuff that we've been working on is in industry like manufacturing industry and stuff like that. 私たちが取り組んできたことの多くのアプリケーションは産業用です。 Obviously, you're you're kind of pursuing world models and and and in that broader world. 世界モデルやその広い範囲を追求されていますね。 And I think there's other people that have come at the world model pace from a more like generative approach. より大きなスケールでの動画モデルから世界モデルに取り組んでいる人たちもいますね。 And so I think you've got folks, you know, uh got the Google folks in Genie and the video models. Google の Genie や、ロボット分野の方々などがいます。 You've got folks, you know, building VAS on the robotic side. ロボット側で VLAs を構築している人たちもいますよね。 You've got uh FE and and kind of like the 3D spatial models. FE や 3D 空間モデルのような取り組みもありますね。 as you think about kind of the the body of of of of evidence that got you excited about the JEPA models and how you kind of compare them to what the generative folks have done, you know, where do you think we are today in in terms of like comparing these architectures and approaches? あなたを惹きつけた証拠を振り返ると、どんなものが浮かびますか？ Okay, so what model is quickly becoming a buzzword? 世界モデルはすでに流行語になりつつあります。 Yeah. そうですね。 Right now, right, certainly in research, but also in industry to some extent. 今は確かに研究の世界でも、ある程度は産業界でも注目されています。 And uh and then there are two factions if you want. 二つの派閥があると言ってもいいでしょう。 I'm not going to talk about VA because VA is clearly now being seen as not going anywhere like it's really not working. VLA については、今やうまくいかないと見なされてきているので触れません。 So VA is you know vision language action models right? VLA というのはビジョン言語アクション モデルのことで、 to basically use the LLM technology to train a system to produce actions for like controlling a robot or something like this, right? LLM の技術を使って、ロボットなどのシステムに行動を生成させるものです。 So you have vision in, language in, action out, maybe language out too. 視覚入力、言語入力、行動出力、そして言語出力もある場合があります。 Um, and that's pretty much now seen as a failure. それはほぼ今や失敗と見なされています。