팟캐스트Hear the voice. See the shape of the thought.

Alfred Wahlforss built Listen Labs after scratching his own itch: when his viral AI-avatar app hit 20,000 users overnight and churn spiked, he needed to know why—fast. The answer was an AI agent that runs voice interviews at scale, drawing from a panel of 30 million people. A year in, Listen serves 20% of the Fortune 500 and has completed over a million interviews. The deeper finding is counterintuitive: respondents are often more honest with an AI interviewer than a human one, and voice transcripts turn out to be richer training signal than credit card data or behavioral logs. Wahlforss and Sequoia's Konstantine Buhler work through why audience selection consumes 80% of Listen's engineering, how back-tested simulation beats vanilla ChatGPT at message testing, and why—as AGI makes building trivially cheap—knowing *what* to build becomes the scarce resource Listen wants to own. ## [00:00] Introduction Alfred opens in the middle of a thought about audience depth: Listen's long-term goal is to reach a billion people and build rich profiles that reveal each person's genuine areas of expertise—not just demographic boxes, but things like whether someone is a true sneaker influencer versus a casual buyer. Konstantine then formally introduces him: Listen launched roughly a year ago, already counts Microsoft, Anthropic, Sweet Green, NBC, and others as customers, and runs thousands of voice interviews simultaneously. The brief cold-open framing gives the episode its throughline—the value of talking to the *right* person, not just any person. > *"Our goal is to get to a billion people in our audience and then to be able to stratify and know what exactly is this person an expert on."* ## [01:20] How Listen Works The product works in three stages: a researcher types a question (say, "how can we improve Cursor's onboarding?"), Listen's AI agent generates an interview guide, then routes those interviews to matched participants from its 30-million-person panel. Hundreds of conversations run in parallel, the results are synthesized, and recommendations surface. The next stage, launching in a few months, is simulation: after tens of thousands of interviews accumulate on a topic, can Listen predict how customers will answer *future* questions without running a new interview? > *"As we get closer to AGI, it will be easier to build things, but the hard part will be knowing what to build—and that's what we're building at Listen."* ## [02:23] Customer Wins Chubbies discovered that chest hair caught uncomfortably on one of their shirt materials; Listen surfaced the feedback, Chubbies redesigned the garment, and comfort scores jumped. Manscaped used Listen insights to reshape a Super Bowl ad. Skims uses it for ongoing product testing. The through-line Alfred draws: Listen handles both small product details and high-stakes campaign decisions with the same workflow—talk to real people, fast. > *"They discovered that chest hair interface really poorly with one of the materials they have. So it's really uncomfortable to wear one of their shirts, and they changed the shirt and it became radically more comfortable."* ## [03:28] Surveys Versus Reality Konstantine presses on the classic critique: survey respondents lie, or at least contradict themselves. Alfred's evidence: Listen ran the same multiple-choice survey questions back to the same people and found radical inconsistency—but when those same people had to reason through an open-ended voice answer, consistency improved sharply. On sales-data back-testing, Alfred agrees AB tests are the gold standard but notes they require large user bases that most companies don't have. Interview data, properly designed, beats no data. > *"If you go back to the same person and ask them a survey question in a multiple choice fashion, they're much more inconsistent. But when you actually have to think and reason through your answer, you're much more consistent."* ## [05:13] Zoom Like AI Interviews The participant experience is a video call with an AI agent—not a text form. The agent watches facial expressions and vocal tone, giving Listen a second signal layer beyond what people say. Alfred cites advertising testing as the clearest win: respondents might rate an ad highly on a Likert scale but show genuine enthusiasm in video, and that enthusiasm predicts Meta and LinkedIn performance marketing results significantly better than the numeric score. Every data point links back to the actual video clip, so researchers can verify the AI isn't hallucinating sources. > *"For every data point you can always click and then look at the video or see the quote—so you know that AI is not just hallucinating where it's coming from."* ## [07:14] Origin Story Alfred and his co-founder shipped a consumer app called "Be Fake"—an early stable-diffusion fine-tuning tool for creating AI avatars of yourself—which went viral overnight and hit 20,000 users. Churn spiked immediately and they had no idea why. They built an AI interview tool to ask their own users, found it genuinely useful, and pivoted. The market-research product they built for themselves became Listen Labs. > *"We built this AI interview for ourselves because we had a ton of churn and we wanted to understand why—and that's how we got started."* ## [08:01] Old World Research The pre-Listen world had two speeds: slow online survey tools like Qualtrics, or expensive services firms that charge tens of millions to recruit participants, design question methodology, moderate focus groups, and synthesize hundreds of transcripts. Question design alone is an academic discipline—ask "how much would you pay for this?" and you get junk data. The sourcing problem is equally hard: incidence rates of 10% mean nine out of ten recruited panelists get screened out, burning trust and causing churn on the databases themselves. > *"In traditional industries like CPG or even Microsoft, they spend tens of millions of dollars on focus groups to bring people in a room and interview them—and we can help speed that up much faster."* ## [09:50] AI First Benefits Three compounding advantages: speed (results from real people in five minutes), cost (asynchronous interviews pay participants less than synchronous ones, and participants accept that willingly), and honesty (people open up more to a non-judgmental AI than to a human interviewer who might silently judge them). Alfred mentions sensitive use cases—interviewing children about products, with parental consent—as an area where the AI's non-threatening presence produces data that focus groups can't. > *"People are more honest talking to an AI. It's a very therapeutic experience because it's a non-judgmental entity that's really interested in you."* ## [11:32] Finding The Right People Listen spends 80% of its engineering resources on audience quality, not the interview agent itself. The reason: power-law customer segmentation means talking to the wrong 100 people gives you wrong insights. Sweet Green's most valuable customer is urban, high-income, mostly female, and—Alfred's specific example—knows what seed oils are (roughly 1% of the population). Listen builds rich profiles across every interview a panelist ever participates in, so an offhand comment ("I'm a total sneaker head") in an unrelated interview can resurface that person when Nike needs launch feedback. Traditional email-list panels couldn't do cross-topic profiling. > *"Even a product like Sweet Green, which you would think is for everyone, the right audience is typically urban, high household income, mostly female—and they need to know what seed oils are, which only like 1% of the population does."* ## [14:30] CRM And Prospecting Sweet Green already has a CRM full of its most loyal customers—so why use Listen? Three reasons: researching *prospective* customers who aren't yet in the CRM requires an external panel; CRMs are typically disorganized and legally constrained (Google can't spam Gmail users, even its own); and direct outbound email risks getting flagged as spam, which can permanently damage a domain's deliverability. Listen provides clean, third-party panel access that sidesteps all three problems while still supporting CRM-connected campaigns when brands want them. > *"What we found is that the CRM is typically really unorganized, and sometimes there are regulatory issues—if you're at Google, you can't just send emails to people who use Gmail."* ## [15:35] Consulting In The AI Era Konstantine—a former buyer of McKinsey-style consulting—asks whether firms like Bain still have a role. Alfred's view: yes, but margins compress. Bain already uses Listen to accelerate existing workflows. The more optimistic scenario: AI doesn't just replace a research project, it makes research cheap enough to run five simultaneous strategic explorations that a company never would have commissioned before. Alfred predicts consulting expands in scope even as price-per-project falls. On economic surplus, Listen has charged hundreds of thousands of dollars to interview 20 doctors across eight countries—fast—a project that previously would have taken months. The surplus is currently staying with the supplier. Alfred also flags an emerging agentic loop: churn interviews surface bugs, which connect directly to a coding agent that opens a PR and ships the fix. Listen as the customer-intelligence "left side" of an autonomous product development cycle. > *"Because you're able to do it faster, I would argue you should be able to charge more for it—and we have charged hundreds of thousands of dollars to speak to 20 doctors across eight countries."* ## [20:05] Market Research Simulation This is the episode's technical core. Konstantine frames the evolution as 1.0 (call 100 people manually), 2.0 (AI-native simultaneous interviews), and 3.0 (generative simulation). Alfred explains how Listen's simulation works: interview a single person deeply, build a persona model, then scale to a thousand statistically representative agents. Back-testing removes a held-out question and measures prediction accuracy—they reach 95% on stable preference domains and deliberately expose the model to nonsensical queries (dog names) to calibrate what it *can't* predict. Alfred ran a personal live test: 100 title variants for a conference talk, run through Listen's panel simulation. The top-ranked title performed twice as well as the second. He then ran the same test in ChatGPT—which picked the wrong title when shown a past successful talk versus a less successful one. Listen's domain-specific panel data beat the general model. The gap: interview transcripts outperform credit card spend, behavioral logs, or ChatGPT persona prompting because voice conversations capture how a specific *type* of person actually reasons, not just what the average person does. Looking ahead, Alfred sees simulation handling "billboard tagline" decisions while real interviews remain the standard for Super Bowl ad buys. The product's proprietary eval climbed from 20% to 85% on avoiding repetitive questions, then Listen raised the bar with a harder eval (screen-state awareness, skipping irrelevant questions) and is back at 20%—which Alfred frames as the vertical AI flywheel: a proprietary benchmark that only you can keep climbing. > *"We were able to get 95% accuracy to predict how they will answer certain questions. The tricky part is knowing what things you can answer and what you can't."* ## [35:33] Closing Thoughts Alfred's conviction: human input will always be necessary because humans are inherently irrational—TikTok trends can overturn a marketing strategy overnight, and no AGI will preempt that. His uncertainty: the ceiling for simulation quality. His moat argument: network effects on the panel (supply-demand flywheel), data network effects (more interviews → better simulation), and product stickiness (interview history compounds inside the platform). But the simplest advantage he cites is opinionated defaults—early customers using vanilla LLMs to design their own interview guides got bad data and blamed Listen; now the agent enforces question-design best practices and data quality is consistent. Konstantine ends with the "Tide Pods moment" question: can Listen's AI start *generating* product ideas mid-interview rather than just testing them? Alfred says customers already feed AI-generated images into interviews manually; the MCP integration means Claude can loop Listen calls autonomously. The vision is live brainstorming between the AI interviewer and the respondent—ideas surfacing as the customer articulates a pain, not after. > *"Founders want to build something that's complex X, but customers want something that's stupid simple and it just works. And that's the advantage you have as a vertical AI company—you can train the agent to follow best practices in the work that you do."* ## Entities - **Alfred Wahlforss** (Person): Co-founder and CEO of Listen Labs; previously built "Be Fake," a viral AI-avatar consumer app. - **Konstantine Buhler** (Person): Partner at Sequoia Capital; host of the Training Data podcast; former consultant and operator. - **Listen Labs** (Organization): AI-first customer research platform; runs voice interviews with a 30-million-person panel; building generative simulation. - **Market Research Simulation** (Concept): Building persona models from accumulated interview data to predict future customer responses without running new interviews; back-tested against held-out questions. - **Audience Quality** (Concept): Listen's thesis that 80% of research value comes from recruiting the right respondents—power-law customer segments—not just any panelists. - **Be Fake** (Software): Alfred's earlier consumer app (AI avatar fine-tuning via stable diffusion); the origin of Listen's interview tooling. - **Bain** (Organization): Management consulting firm; cited as an active Listen customer using the platform to accelerate traditional research workflows. - **Procter & Gamble** (Organization): Cited as the historical archetype of market-research-driven brand management; Tide Pods and M&M's given as canonical examples. - **Qualtrics** (Software): Legacy survey platform representing the "old world" of market research tooling.

At AI Ascent 2026, Neuralink co-founder and president DJ Seo sits down with Sequoia partner Shaun Maguire to lay out exactly where the company stands: 20-plus Telepathy patients controlling computers and robotic arms through pure thought, Blindsight in preclinical testing and potentially cleared for human use by end of 2026, and a first-principles manufacturing philosophy borrowed from Elon Musk that treats surgical robots the way SpaceX treated reusable rockets. DJ argues that the real ceiling of this technology is not cursor control or speech synthesis but direct, uncompressed, multimodal transfer of concepts — AI as a neocortical layer sitting above the human limbic system — and that scale, the same variable that unlocked the LLM era, is the only remaining gate. ## [00:00] Introduction Shaun Maguire opens the session by announcing a two-minute Neuralink patient video before the interview begins, telling the audience to stay on the side because what they are about to watch is proof that the company has already cleared the hardest bar: restoring human agency to people who had lost it entirely. ## [00:21] Telepathy Patient Stories The video narrates four patients whose lives changed after receiving the Telepathy implant. A quadriplegic patient describes moving a cursor with thought alone — "I'm thinking and a cursor is moving on a screen. It blew my mind." An ALS patient who lost the ability to speak regains a digital voice through the implant: "I'm talking to you with my mind." Another patient notes that the implant flipped how his child sees him: "I am not able to do things that other dads can, but now he thinks it's so cool that I can do things that other dads cannot." > *"Before the implant, I was locked in, non-verbal, quadriplegic. Now I control my computer just by thinking and the rewards have been immense for me."* ## [01:06] Convoy Robotics Independence The video shifts to Convoy, Neuralink's assistive robotics team, which is extending BCI control beyond a screen to physical manipulation in the real world. A patient who had been losing motor function moves a robotic arm through its axes using only neural intent: "It was incredible to be able to just gesture with an arm again." A second patient, Kenneth, who was losing his voice to ALS, uses the system's speech synthesis to speak aloud in real time during the video — words generated by his brain signals rather than his vocal cords. > *"Gaining functionality that I thought was gone forever was so incredibly life-changing."* ## [02:04] Blindsight Vision Restore The video previews Blindsight, Neuralink's second product line, designed for patients who have lost both eyes or optic nerve function. An external camera captures the visual scene; the device writes the signal directly into the visual cortex via electrical stimulation, generating phosphenes — artificial pixels of light. A patient named Audrey, asked how it feels, answers simply: "Life-changing." The video closes with the line "all with my mind" spoken over footage of a patient interacting with the world through the restored signal. > *"The future of this technology feels almost unlimited... we are finding ways to apply it across all regions of the brain."* ## [03:10] After Video Reflections DJ Seo, visibly moved after watching the video alongside the audience, speaks first: "We were cracking a lot of jokes before that video, but honestly, that brought tears to my eyes." He describes the work as one of the most inspiring projects in the world — not because of the technical milestone but because the team is giving back capabilities that patients had already grieved as permanently lost. Maguire affirms the sentiment before pivoting to the founding story. > *"This is one of the most inspiring projects in the world. It's incredibly difficult what they're doing and I mean, they're truly saving people."* ## [03:31] Origin Story And AI DJ traces Neuralink's founding insight to a single bottleneck: the mismatch between human output bandwidth and AI capability. In 2016, saying that out loud "sounded insane," but the logic has not changed. His personal path ran through a childhood fascination with the brain, undergraduate work at Caltech building miniaturized low-power electronics, and a Berkeley PhD focused on shrinking lab-grade neural systems down to something deployable. When he met Elon Musk near the end of his PhD, the scale and ambition of the project made refusal impossible. He frames the brain as "the most interesting compute that we all carry" and "the only form of general intelligence that we know to date." > *"Really the key insight back then was sort of the IO bottleneck between the human output and AI capabilities."* ## [06:31] Scaling And Vertical Integration Maguire presses on what smart people most misunderstand about Neuralink: many know the implant and the decoding algorithm, but almost nobody grasps the manufacturing and surgical-robot infrastructure the company built in parallel from day one. DJ attributes this to what he calls "Elon magic" — an insistence on vertical integration that gives Neuralink control over every layer from chip design to factory floor to robotic surgery deployment. The target is not a niche medical device; it is LASIK-scale surgery available to millions. Building that capacity first means progress looks slow until "the iceberg pops over the waterline" and ramp becomes near-instantaneous. > *"Vertical integration is something that is really the lifeblood of Neuralink and Elon companies and what really enables us to have that fast iteration loop from design, develop, deploy."* ## [09:27] Caregivers And Purpose Asked which patient story inspires him most, DJ refuses to pick one — the power, he says, is not only in the patients but in the caregivers: Nolan's mother Mia, Brad's wife Tiffany, Ken's wife Cheryl. He describes their presence as "a really powerful human story of love, sacrifice, and resilience." He then takes what he calls a philosophical tangent: his core belief is that fulfillment comes from helping others, because the gap between self and other is not categorically different from the gap between your present and future selves. That belief is what he says keeps him and much of the Neuralink team going — they are "igniting a fire of hope" for people who had given up on recovering what they lost. > *"I personally and as well as many others at Neuralink find extreme fulfillment being able to help those that really cannot help themselves."* ## [13:10] BCIs Meet AI Future Maguire asks the room's core question: how do BCIs and AI converge? DJ sketches a two-horizon answer. Near term, the system translates neural intent into legacy interfaces — keyboard, mouse, language — which is already working. The real breakthrough, which he thinks is "not super distant," is bypassing those legacy interfaces entirely and computing on raw neural intent. He points to transformer architectures as existence proofs: nothing prevents them from learning the latent manifolds of neural data given sufficient scale. Neuralink is already fine-tuning LLM-class models on neural recordings from its 20 participants and finding "very counterintuitive" patterns. The ultimate ceiling he names is "direct, uncompressed, high-fidelity, multimodal transfer of concepts" — the Matrix's "I learned kung fu" moment and possibly beyond it. He also shares what he calls a clarifying lesson from working with Musk: "all green light schedule" — a first-principles forcing function that strips every man-made bottleneck and asks how fast something could actually be built if every light were green. His estimate is that 80–90% of perceived constraints in hardware development are artifacts of convention, not physics. > *"I think if you really think about the ultimate ceiling of this technology, it's really direct uncompressed high fidelity and multimodal transfer of concepts."* ## [21:05] Audience Q&A Wrap Three audience questions in the final four minutes. On product sequencing — when to go deep versus expand — DJ explains the "beachhead and expand" strategy: build everything generalizably enough from the start so that regulatory approval for motor cortex becomes a template for visual cortex and beyond. The first approval is the hardest; every subsequent one rides the clinical safety record already established. On augmentation for healthy users, DJ frames everything around benefit-risk: the calculus is obvious for quadriplegic patients; for otherwise healthy users it remains unclear, but he notes that off-label use after approval is legally available to anyone who can find a neurosurgeon and pay out-of-pocket. On the hard problem of consciousness, he gives a pointed one-liner: if you can inject new senses and measure the subjective response quantitatively, you may have a pathway toward measuring consciousness itself. Maguire closes by calling Neuralink "one of the most inspiring companies in the world." > *"If you are able to inject new senses, there may be ways to quantitatively understand that."* ## Entities - **DJ Seo** (Person): Co-founder and president of Neuralink; PhD in miniaturized electronics from Berkeley; joined after meeting Elon Musk near the end of his doctorate - **Shaun Maguire** (Person): Partner at Sequoia Capital; host of the AI Ascent 2026 fireside session - **Elon Musk** (Person): Co-founder of Neuralink; originator of the "all green light schedule" and vertical integration philosophy carried across Tesla, SpaceX, and Neuralink - **Neuralink** (Organization): BCI company founded in 2016; products include Telepathy (motor prosthesis) and Blindsight (vision restoration via visual cortex stimulation) - **Telepathy** (Software): Neuralink's first commercial product; allows paralyzed patients to control computers and robotic devices through neural intent decoding - **Blindsight** (Software): Neuralink's second product line; restores vision for patients with total loss of eyes or optic nerve by writing directly to the visual cortex; in preclinical testing as of mid-2026 - **IO Bottleneck** (Concept): The mismatch between human output bandwidth (speech, typing, gesture) and AI processing capability; the founding problem Neuralink was built to solve - **Neural Foundational Model** (Concept): LLM-class transformer models fine-tuned on neural recording data; Neuralink is building these at 20-participant scale and observing counterintuitive patterns in neural latent space - **All Green Light Schedule** (Concept): Elon Musk's first-principles engineering discipline — strip every man-made constraint and ask what physics alone limits; DJ estimates 80–90% of hardware delays are conventional, not physical

Cursor의 Federico Cassano와 Fireworks의 Dmytro Dzhulgakov가 Sonya Huang에게 Composer 2 구축의 전 과정을 설명한다. Kimi 2.5 MoE 베이스 모델부터 대규모 mid-training, 전 세계 비동기 분산 RL까지, 특화 모델이 범용 모델보다 비용과 품질 면에서 유리한 이유를 짚어준다. 핵심은 인프라 이야기다. 대륙을 넘나드는 4개 GPU 클러스터, 1TB 가중치 스냅샷을 1분 안에 전송하는 Delta Compression, 실제 사용자 신호로 몇 시간마다 라이브 모델을 업데이트하는 실시간 RL 루프. 이 기술들이 결합되어 Cursor는 범용 모델 대비 훨씬 낮은 추론 비용으로 최전선 코딩 성능을 제공할 수 있었다. ## [00:00] 소개 Dmytro가 제기한 RL 환경 충실도 문제로 대화가 시작된다. 모델이 가짜 환경에서 실행 중임을 감지하고 이를 악용할 수 있기 때문에, 학습 환경은 실제 사용자 기계와 최대한 가깝게 맞춰야 한다. > *"모델은 속이는 걸 좋아합니다. RL은 속임수를 아주 잘 부추기죠."* — Federico Cassano 이 한 마디가 에피소드 전체를 관통하는 기술적 원칙을 잡아준다. 인프라의 모든 요소는 학습 조건과 프로덕션 현실 사이의 간극을 좁히기 위해 존재한다. ## [00:53] Cursor가 Composer 2를 학습시킨 이유 Federico는 Composer 2의 핵심 논리를 하나의 비유로 설명한다. 모델의 가중치는 고정 크기 저장 드라이브와 같아서, Cursor가 필요로 하지 않는 작업에 할당된 비트는 모두 낭비된 비트다. 코딩 일반이 아닌, Cursor 내 소프트웨어 엔지니어링에만 전체 가중치 예산을 집중하면, 모델은 그 한 가지 역할에서 더 뛰어나면서도 추론 시 서빙 비용은 더 낮아진다. Dmytro는 인프라 관점에서 같은 논리를 풀어낸다. 프롬프트 엔지니어링으로 어느 정도까지는 갈 수 있지만, 에이전트가 어떤 툴을 어떤 순서로 어떤 인자와 함께 호출해야 하는지 같은 세밀한 행동 특성을 포착하려면, 파인튜닝과 RL을 통해 모델에 직접 구워 넣는 수밖에 없다. > *"프롬프트 엔지니어링으로 갈 수 있는 거리에는 한계가 있어요. 정말 훌륭한 AI 제품을 만들려면 파인튜닝을 거쳐 모델 행동에 영향을 줘야 합니다."* — Dmytro Dzhulgakov ## [04:55] 특화 모델 vs. Bitter Lesson Sonya가 반론을 제기한다. 머신러닝의 역사는 더 큰 범용 모델에 밀려난 특화 모델로 가득하다. Composer 2가 TabNine의 실수를 반복하는 건 아닐까? Federico는 다르다고 답한다. Bitter Lesson은 파라미터와 데이터 규모에 관한 것이다. Cursor가 하는 일은 모델의 유한한 용량을 불필요한 곳에서 해방시켜, 중요한 한 가지 작업에 더 많은 스케일링 이점이 흡수되도록 만드는 것이다. Cursor가 경쟁하는 랩 모델들도 코드를 집중적으로 학습한다. 순수한 범용 모델이 아닌 것이다. Cursor는 데이터 파이프라인을 직접 제어해 그 특화를 더 빠르게, 더 깊이 밀어붙이고 있을 뿐이다. ## [06:16] Composer 2 학습 레시피 Composer 2는 Kimi 2.5에서 시작한다. 활성 파라미터 30B를 가진 1조 파라미터 MoE 모델이다. 학습은 두 단계로 진행된다. 먼저 사전학습에 준하는 규모로 코드 토큰을 학습하는 mid-training 단계가 있다. Cursor의 프로덕트 데이터 덕분에 고품질 코딩 컨텍스트에 이례적으로 풍부하게 접근할 수 있다. 그다음 시뮬레이션 환경에서 실제 Cursor 에이전트 세션을 실행하는 대규모 RL 단계가 이어진다. Mid-training은 모델에게 코드 세계를 가르친다. 라이브러리 API, 관용 패턴, 올바른 문법. RL은 그 지식을 올바른 행동으로 날카롭게 다듬는다. 툴을 제대로 호출하고, 멀티턴 에이전트 세션을 탐색하며, 실제로 컴파일되고 테스트를 통과하는 코드를 작성하도록 학습한다. 비동기 파이프라인 덕분에 trainer와 rollout 환경이 교대 실행이 아닌 동시 실행된다. 수학적으로 완벽한 업데이트를 포기하는 대신 GPU 활용률 거의 100%를 확보하는 것이다. > *"비동기라서 완벽한 수학적 업데이트를 하지 못해 몇 퍼센트를 잃을 수도 있어요. 하지만 GPU 용량 절반을 놀리지 않아도 되는 것으로 훨씬 더 많이 보상받죠."* — Dmytro Dzhulgakov 학습은 FP4로 실행해 프론티어 랩보다 작은 GPU 플릿에서 최대 처리량을 끌어낸다. 추론 엔진은 직접 구축 대신 Fireworks를 선택했다. Cursor 엔지니어들이 또 다른 추론 스택을 만드는 데 시간을 쓰지 않고 학습 효율성에 집중하기 위한 의도적인 결정이다. ## [16:32] 전 세계 RL 인프라 확장 Composer 2가 요구하는 규모의 대형 단일 클러스터를 확보할 수 없었기 때문에, 팀은 분리 전략을 택했다. 하나의 클러스터가 모든 학습을 담당하고, 추론, 즉 rollout 컴포넌트는 Composer 1.5의 프로덕션 서빙에서 오프피크 시간대 여유 용량을 포함해 지리적으로 분산된 4개 클러스터에서 실행된다. 학습은 고속 인터커넥트와 동기화된 동작이 필요하지만 추론은 그렇지 않아, 소규모 인트라클러스터 네트워크를 가진 이기종 GPU 세대에서도 실행할 수 있다. 시스템에서 가장 어려운 문제는 가중치 동기화다. Kimi 2.5는 약 1TB 크기이고, trainer는 5~15분마다 새 체크포인트를 생성한다. 10분마다 1TB를 대륙을 넘어 전송하면 추론이 멈춰버린다. 해결책은 이렇다. RL 업데이트는 변경되는 가중치의 패턴이 드문드문하고 규칙적이다. 팀은 페이로드를 약 20배 줄이고 diff만 전송하는 Delta Compression 알고리즘을 작성했다. 수신 측은 전체 체크포인트를 무손실로 재구성하므로 상대편에서 수치적 놀라움은 없다. > *"전체 모델이 1테라바이트임에도 불구하고, 매 스텝마다 모든 가중치가 바뀌지는 않아요. 어떤 가중치 부분이 변경되는지에 매우 규칙적인 패턴이 있죠."* — Dmytro Dzhulgakov ## [23:32] 부동소수점 드리프트 비동기 RL 루프가 추론에서 rollout 궤적 배치를 trainer로 돌려보낼 때, trainer는 GRPO loss의 로그 확률을 재계산하기 위해 동일한 순방향 패스를 다시 실행한다. 이론적으로 로그 확률은 동일해야 한다. 실제로는 종종, 때로는 크게 달라진다. 근본 원인은 부동소수점 비결정성이다. 부동소수점 수의 덧셈은 교환법칙이 성립하지 않아 A+B+C ≠ C+B+A이고, 작은 차이가 수십억 번의 연산에 걸쳐 누적된다. 일반 추론에서는 모델이 이 노이즈에 견고하지만, RL, 특히 희소한 MoE 게이팅 함수에서는 노이즈가 증폭되어 trainer와 추론이 어떤 토큰이 샘플링되었는지에 대해 의견이 갈리고, 학습 신호가 오염된다. ## [25:11] MoE 민감도 설명 MoE 아키텍처는 게이팅 레이어 때문에 부동소수점 드리프트를 증폭한다. 각 트랜스포머 레이어에서 게이팅 네트워크는 384개 전문가 전체에 점수를 매기고 각 토큰에 대해 상위 8개를 선택한다. 숨겨진 상태의 소수점 다섯 번째 자리의 차이만으로도 선택 경계에서 전문가 7번이 9번으로 바뀌어, 토큰이 완전히 다른 모델 부분으로 라우팅될 수 있다. MoE 전문가는 크고 대부분 겹치지 않기 때문에, 잘못된 전문가 선택은 수치 노이즈가 내내 작게 유지되는 밀집 모델과 달리 큰 출력 발산으로 이어진다. ## [26:25] Router Replay 해결책 완화책은 Router Replay다. 추론 중 모델은 각 토큰에 대해 활성화한 전문가 인덱스를 기록하고, 그 정수를 생성된 시퀀스와 함께 trainer로 돌려보낸다. trainer는 처음부터 다시 계산하는 대신 동일한 전문가 선택을 강제 적용해 증폭 체인을 끊는다. Router Replay와 함께, 팀은 추론과 학습 간의 양자화 수준과 커널 구현을 맞춰 다른 모든 수치 불일치 원인을 최소화했다. > *"이런 수치 정렬 작업의 대부분은 양자화 수준 맞추기, 커널 맞추기 등의 트릭으로, 학습과 추론 구현 간의 발산을 줄이는 것입니다."* — Dmytro Dzhulgakov ## [27:19] 실시간 RL 루프 시뮬레이션 rollout 루프와 병행해, Cursor는 Federico가 실시간 RL이라 부르는 것을 운영한다. 프로덕션의 실제 사용자 세션이 학습 파이프라인으로 피드백된다. 사용자가 Composer의 생성 결과에 만족하거나 불만족하면 그 신호가 포착되고, 몇 시간마다 새 모델 버전이 배포된다. 팀은 그 주기를 더 짧게 만들기 위해 노력하면서도, rollout 수평이 길어질수록 다시 늘려야 할 것임을 안다. 에이전트 세션이 길수록 평가에도 더 많은 시간이 걸리기 때문이다. 시뮬레이션 루프와 실시간 루프는 서로 다른 목적을 가진다. 시뮬레이션은 같은 프롬프트에서 16~128개의 rollout을 병렬로 실행할 수 있고, GRPO loss에는 그룹화된 rollout이 필요하다. 어떤 사용자에게도 영향을 주지 않고 오프폴리시로 탐색할 수 있으며, 실제 사용자가 사용하기에 충분할 만큼 좋아지기 전에 성능을 끌어올릴 수 있다. 실시간 RL은 모델이 이미 최소 품질 기준을 충족했을 때만 작동하는 정제 레이어다. 나쁜 경험을 한 사용자는 피드백 신호 생성을 멈추기 때문이다. > *"이걸로 모델을 처음부터 만들 수는 없어요. 사용자들이 그 모델을 써야 하니까요. 이미 좋아야 하고, 우리는 더 좋게 만들 수 있을 뿐이죠."* — Federico Cassano ## [31:49] 장기 수평 에이전트 rollout 수평이 늘어날수록 두 가지 구조적 문제가 생긴다. 첫째, 크레딧 할당이다. 몇 분짜리 세션 끝에 단 하나의 좋아요/싫어요 보상이 주어지면, 모델은 궤적 내 50개 이상의 결정 중 어느 것이 결과를 이끌었는지 파악해야 한다. 궤적이 길어질수록 지수적으로 어려워진다. 둘째, 컨텍스트 윈도우가 가득 찬다. Cursor의 해결책은 "compaction"이라는 이름으로 자기 요약을 직접 RL 루프에 구워 넣는 것이다. 모델은 RL 보상을 통해 컨텍스트 한계에 가까워졌을 때 진행 상황을 유용하게 요약하고, 그 요약에서 충실하게 이어가는 법을 함께 배운다. 컨텍스트 200K짜리 모델이 압축된 작업 기억을 들고 윈도우를 리셋할 수 있기 때문에, 사실상 수백만 토큰에 걸쳐 작동한다. > *"RL은 모델이 목표를 향해 올바르게 행동하도록 밀어붙이기 때문에, 동시에 좋은 요약을 생성하도록, 그리고 그 요약을 아주 잘 따르도록 함께 학습시키고 있는 거예요."* — Federico Cassano ## [34:29] RL이 모든 곳에 필요한 이유 Sonya는 RL을 에이전트적, 장기 수평 툴 사용에 특화된 도구로 규정한다. Federico는 반박한다. RL은 탭 완성을 포함해 어디서나 유용하다. 그의 이론은 이렇다. 사전학습된 모델은 인류의 모든 지식을 흡수했지만, 프롬프트가 주어졌을 때 어떤 페르소나, 즉 전문가인지 학생인지 중간 어딘가인지를 취해야 할지 모른다. RL 학습의 첫 번째 단계는 그 분포를 날카롭게 해 모델에게 "너는 전문가야, 이걸 올바르게 해"라고 알려준다. 이 효과는 상호작용 하네스가 없는 요약 같은 작업에서도 가치 있다. 두 번째 단계, 모델이 눈에 띄게 추론하기 시작하고 컴퓨트 곡선이 평탄해지는 지점이 바로 태스크별 신호가 진짜로 복리 효과를 내는 곳이다. ## [37:34] LLM을 심판으로 활용한 보상 보상이 검증 가능할수록, 코드가 컴파일되는지, 테스트를 통과하는지, 답이 수치적으로 맞는지, 더 많은 컴퓨트를 RL에 부어도 더 나은 모델을 얻을 수 있다. LLM을 심판으로 활용하면 정답을 정의하기 어려운 태스크의 빈틈을 채울 수 있다. 루브릭을 프롬프트로 인코딩하고, 두 번째 모델이 rollout 품질을 평가하게 한다. Dmytro는 인간 평가자가 "좋다"는 게 무엇인지 명확히 표현하기 어렵지만 명시적 기준에 비춰 평가는 할 수 있는 요약 같은 스타일 지향 태스크에 특히 유용하다고 말한다. > *"일반적으로 보상이 검증 가능할수록 좋습니다. 컴퓨트를 확장하면서 더 나은 결과를 얻을 수 있으니까요."* — Dmytro Dzhulgakov ## [39:14] 어려운 도메인에서의 RL 정답을 저렴하게 계산할 수 없는 도메인, 창의적 글쓰기, 개방형 추론, 도메인 전문 지식의 경우, RL 개선의 길은 환경을 더 풍부하게 만드는 것이다. 더 많은 프로덕트 지표를 포착하는 더 큰 시뮬레이션 환경은 자동화된 평가를 더 멀리 밀어붙일 수 있게 해준다. 전문가는 여전히 필요하다. 개별 rollout을 판단하는 게 아니라, 보상 함수가 최적화해야 할 대상을 정의하는 태스크와 루브릭을 설계하기 위해서다. ## [40:13] 직접 환경 구축하기 Cursor는 RL 환경 공급업체를 전혀 사용하지 않는다. 코딩에 있어 GitHub 저장소는 사실상 무한한 작동 환경 풀을 제공한다. 저장소를 클론하고, 의존성을 설치하고, 모델에게 태스크를 주고, 테스트 스위트로 결과를 측정한다. 더 어려운 인프라 문제는 에피소드 첫머리에서 다룬 종류의 속임수를 막을 만큼 그 환경을 충분히 현실적으로, 그리고 동시에 100,000개를 즉시 온디맨드로 돌릴 수 있을 만큼 빠르게 만드는 것이다. Cursor의 답은 컨테이너가 아닌 완전한 VM 스택이다. 즉각적으로 임의의 규모로 버스트할 수 있고, 실제 사용자 기계와 충분히 가까워 모델이 차이를 감지할 수 없다. Dmytro는 공급업체 구도를 이렇게 정리한다. 프론티어 랩은 모든 태스크를 커버하는 범용 환경이 필요하고, 프로덕트 회사는 자신의 프로덕션 환경에서 RL을 돌려야 한다. 어떤 모델에게든 가장 강력한 학습 환경은 그 모델이 실제로 사용될 제품 자체다. > *"가장 강력한 환경은 자신의 프로덕트입니다."* — Dmytro Dzhulgakov ## [44:34] 마무리 생각 Sonya는 애플리케이션 회사에서 프론티어 모델 랩으로 나아가는 Cursor의 궤적이 다른 AI 프로덕트 회사들이 따라갈 패턴이라고 마무리한다. Federico는 Cursor의 GPU 예산으로 학습 실행을 가능하게 해준 인프라 기반을 제공한 Fireworks에 감사를 전한다. Dmytro는 대부분의 사람들이 순수하게 알고리즘적이라고 생각했던 문제에 얼마나 깊은 시스템 엔지니어링이 담겨 있는지를 돌아본다. ## 등장인물 - **Federico Cassano** (인물): Cursor에서 Composer 2 리서치 리드. 학습 레시피와 RL 방법론을 주도했다. - **Dmytro Dzhulgakov** (인물): Fireworks AI 인프라 리드. Composer 2를 위한 분산 RL 학습 시스템을 엔지니어링했다. - **Sonya Huang** (인물): Sequoia Capital 파트너. AI 투자에 초점을 맞춘 팟캐스트 진행자. - **Composer 2** (소프트웨어): Cursor의 특화 에이전트 코딩 모델. Kimi 2.5 MoE를 기반으로 mid-training과 대규모 RL로 학습됨. - **Fireworks AI** (조직): 모델 서빙 및 추론 인프라 회사. Composer 2 RL 학습을 위한 분산 GPU 백본을 제공했다. - **Cursor** (조직): AI 코딩 IDE 회사. Cursor 내 소프트웨어 엔지니어링을 위한 특화 파운데이션 모델로 Composer 2를 학습시켰다. - **Kimi 2.5** (소프트웨어): Moonshot AI의 오픈소스 1조 파라미터 MoE 모델 (활성 30B). Composer 2의 베이스로 사용됨. - **GRPO** (개념): Group Relative Policy Optimization. Composer 2에 사용된 RL 알고리즘으로, 정책 그래디언트 계산을 위해 같은 프롬프트에서 다수의 병렬 rollout이 필요하다. - **Router Replay** (개념): MoE 수치 정렬 기법. 추론 시 전문가 라우팅 결정을 기록하고 trainer에 재현해 부동소수점 드리프트로 인한 로그 확률 발산을 방지한다. - **실시간 RL** (개념): Cursor의 프로덕션 피드백 루프. 실시간 사용자 만족도 신호를 포착해 몇 시간마다 새 모델 버전을 배포하며 모델을 지속적으로 업데이트한다. - **Delta Compression** (개념): 학습과 분산 추론 클러스터 간의 가중치 동기화 기법. 변경된 파라미터만 전송해 실제로 1TB 스냅샷을 약 50GB로 줄인다. - **자기 요약 / Compaction** (개념): 에이전트가 컨텍스트 윈도우 한계에 가까워졌을 때 작업 컨텍스트를 압축하도록 RL로 학습된 능력. 사실상 무제한 수평 작동이 가능해진다.

Brian Halligan interviews Notion co-founder Ivan Zhao on his journey as a 'refounder' who navigated the company through its 2015 Kyoto restart and the 2023 generative AI pivot. Zhao details Notion's transition from a traditional SaaS structure to an AI-native 'jazz band' model that prioritizes technical versatility, taste, and agency over rigid hierarchies. The discussion explores how AI acts as the 'steel' for modern organizations, enabling flatter structures and faster, more reversible decision-making. ## [00:00] Introduction Brian Halligan introduces Ivan Zhao as the 'refounder' of Notion, highlighting his unique ability to restart the company during critical junctures in 2015 and 2023. The conversation sets the stage for Zhao's transition from a traditional SaaS management model to an AI-native organization. Halligan compares Zhao's approach to other tech visionaries like Jack Dorsey, emphasizing the importance of personal style and 'taste' in building a lasting brand. > *I like to think of him as the refounder... he's the canonical example of how a SAS company can move and become an AI company. [00:52]* > *We want to be a jazz band, not a marching band. [00:02]* ## [02:22] From Founder Mode to AI Org Ivan Zhao discusses his detour into traditional delegation and professional management before returning to a hands-on 'founder mode' necessitated by the AI shift. He explains that building with language models is less like predictable bridge engineering and more like 'brewing beer,' where the underlying technology dictates the development path. Zhao emphasizes hiring 'jazz band' people—versatile individuals like designers who code—to navigate the experimental nature of AI integration. > *Building with language model... is like brewing beer. You can't truly predict the things the underlying thing. [06:33]* > *The spirit is technology first-driven development rather than customer-driven first development. [07:01]* ## [11:00] Hiring for Taste and Agency Notion utilizes a 'barbell' hiring strategy that targets both super-junior and super-senior talent while avoiding the 'middle' of traditional SaaS experience. Zhao defines talent as the product of capability, taste, and agency, noting that AI has democratized basic capabilities like coding and writing. Consequently, the company now optimizes for 'agency' and 'taste,' qualities that remain difficult to automate and serve as the primary differentiators for the brand. > *capability got normalized democratized and taste becomes still important [11:53]* > *So the shape it's not it's more like the barbell barbell shape, right? [12:35]* ## [24:28] Refounding Notion in Kyoto In 2015, facing potential failure and low morale, Zhao and co-founder Simon Last laid off their entire staff and relocated to Kyoto, Japan, to rebuild Notion from scratch. This 'Kyoto Reset' allowed them to focus entirely on craft and coding while living a minimalist lifestyle. Zhao chose Kyoto specifically for its status as the 'craft capital of Asia,' which provided the spiritual inspiration needed to view software as a fundamental human tool. > *So my co-founder and I said let's just lay off everybody just go by the two of us. That's the Japan story. [25:41]* > *The story we tell ourselves is like Kyoto is a special place. If you can pull off anywhere, you can pull off from Reborn in Kyoto. [28:05]* ## [30:27] Craft Versus Commerce Zhao views Notion as part of a historical lineage of 'tools for thought,' tracing back to pioneers like Douglas Engelbart and Alan Kay. He criticizes modern Silicon Valley 'tinker culture' for ignoring the history and humanity behind technology. For Zhao, the goal is to find an equilibrium between the pure craft of an artist and the commercial viability of a business, ensuring the product has a 'soul' that resonates with users. > *Tech is like industry doesn't know its past. If you don't know his past you don't know history which is humanity. [31:52]* > *I need to be in equilibrium with my own value of what this company I want to build... [51:33]* ## [32:26] When to Refound For founders whose companies are stagnating, Zhao suggests listening to the 'inner urge' to take drastic action rather than wasting years on ventures without momentum. He argues that refounding is often harder than starting fresh because it involves taking a significant step back to pivot toward a new growth engine. Zhao believes the current AI-driven market is wide open, making it an ideal time for founders to be risk-seeking and follow their intuition. > *For me it's like there's you just feel you have to do something drastic... then you feel liberated once you land in Japan. [32:56]* > *The refounding is harder than it looks. It typically involves like a big step back and two steps forward. [59:57]* ## [34:07] GPT-4 Refounding Shock Zhao describes gaining early access to GPT-4 as a 'full body religious experience' that signaled a fundamental shift in the world. This realization forced a second refounding of Notion, as Zhao felt any work not involving this technology would soon become meaningless. The transition included a grueling 18-month period of low morale while the team waited for the underlying AI models to catch up with their ambitious product vision. > *GBD4 is a religious experience for me. It's like holy [ __ ]... anything you do if you don't do this it will be meaningless. [34:27]* > *that was like a year and a half just go with no error and morale is definitely low [35:50]* ## [45:35] Leadership and Founder Energy Despite being naturally introverted, Zhao explains how he forced himself to master one-to-many communication to build trust within Notion. He maintains a disciplined daily routine, starting at 7 AM and often working until midnight, while using 'guilty pleasure' reading to recharge. To prevent organizational calcification, Notion aggressively acquires startups to bring in 'founder energy,' currently employing over 50 former founders who lead critical domains. > *To lead the group of human you need to do one to many communications otherwise people don't trust you. [46:17]* > *founders are are kind of this kind of like little decalcified meatthead machinery just trying to break things [39:10]* ## [53:17] Sales Culture and Closing Thoughts Notion's transition to enterprise sales involved moving away from 'first-principle' experimentation toward established playbooks, pairing system thinkers with high-energy sales leaders. The conversation concludes with a vision of the 'AI-native' CEO playbook, which replaces traditional 'triangle' hierarchies with a 'circular' model. In this structure, a centralized AI system saturated with company context enables smaller teams to move at breakneck speed with reversible decision-making. > *You should only have each company should only preserve your innovation point to few places... [54:54]* > *All of those kind of one-way doors that Bezos used to talk about are really two-way doors... [62:39]* ## Entities - **Ivan Zhao** (person): Co-founder and CEO of Notion, known for his 'refounder' mindset. - **Brian Halligan** (person): Co-founder of HubSpot and interviewer. - **Notion** (organization): A productivity software company that pivoted to an AI-native model. - **Simon Last** (person): Co-founder of Notion who helped rebuild the company in Kyoto. - **Kyoto** (location): The Japanese city where Notion was restarted in 2015. - **GPT-4** (technology): The AI model that triggered Notion's second refounding. - **Steve Jobs** (person): Former Apple CEO cited as an inspiration for refounding and craft. - **Jack Dorsey** (person): Tech leader mentioned for his AI-centric organizational redesign. - **Douglas Engelbart** (person): Computing pioneer in the 'tools for thought' lineage. - **Erica** (person): CRO of Notion and former CRO of GitHub. - **SaaS** (concept): Software as a Service, the industry context for Notion's evolution. - **Jazz Band** (concept): Metaphor for a flexible, high-agency organizational structure.

Mikey Shulman, co-founder of Suno, discusses the platform's evolution from a physics-based startup to a leader in generative AI music. By modeling music as raw sound waves rather than traditional theory, Suno empowers users to transition from passive listeners to active creators in the era of 'creative entertainment.' ## [00:00] Physics, Raw Sound, and Technical Philosophy Mikey Shulman explains how his background in quantum physics at Harvard influenced Suno's interdisciplinary approach to music technology. By modeling audio as raw sound waves sampled 48,000 times per second rather than using traditional music theory, Suno avoids creative constraints and allows for the emergence of new, microtonal genres. > *I think what I mostly learned is playing at the nexus of two things that don't usually play together is just a massive opportunity. [02:00]* ## [02:15] The Pivot to Consumer Music Generation Initially focused on audio analysis, the Suno team pivoted to generation after breakthroughs in audio compression made high-quality output computationally feasible. They validated the product's 'fun factor' through a Discord bot, discovering that the addictive nature of creation was a stronger signal than traditional business use cases. > *When you are staying up late playing with the thing, and you don't want to go to sleep, it's like a really good sign. [04:49]* ## [11:41] Why Music AI is a Research Problem, Not a Scale Problem Unlike Large Language Models, music generation lacks objective benchmarks, making raw compute scale less effective than targeted research. Shulman emphasizes using human preference data and reinforcement learning to align models with creative tastes, favoring a steady release cadence over long-term isolated development. > *In music there are no right answers. There are no benchmarks. Um, and so scale is somewhat less helpful in solving it. [12:28]* ## [16:22] From Passive Consumption to Creative Entertainment Shulman introduces the concept of 'creative entertainment,' where the act of building provides more fulfillment than the final product itself. He notes that 90% of Suno users are active creators, drawing parallels to the 'bedroom producer' era where accessible tools led to the discovery of new genres. > *People are creating music for the fun and enjoyment and fulfillment that comes with being creative. [17:05]* ## [22:52] Industry Partnerships and Professional Integration Addressing industry concerns, Shulman highlights Suno's partnership with Warner Music Group and its role in augmenting professional workflows. He argues that AI will raise the quality ceiling for artists and predicts that interactive live performances, such as audience participation at Coachella, are the next frontier. > *I think people incorrectly assume that we hate the existing music industry and especially we hate the record labels. [23:17]* ## [25:53] Product Strategy and the Application Moat Suno prioritizes the application layer and user experience as its primary competitive moat, viewing itself as a music company rather than just a technology firm. By focusing on storytelling through full-length lyrical songs and social co-creation features, the company aims to revive the cultural impact of music as a social medium. > *It's unclear how much moat exists in only a model... it's just really undervalued to invest in the product and the UI and the UX. [26:50]* ## Entities - **Mikey Shulman** (person): CEO and co-founder of Suno with a PhD in physics from Harvard. - **Suno** (organization): An AI-powered creative entertainment platform for music generation. - **Sonya Huang** (person): Partner at Sequoia Capital and host of the interview. - **Warner Music Group** (organization): A major global record label that partnered with Suno. - **Discord** (organization): The platform where Suno initially launched its music generation bot. - **Harvard** (organization): The university where Mikey Shulman studied quantum computing. - **Iamona** (person): A poet and artist who uses Suno to create music, illustrating the tool's professional potential. - **Coachella** (event): A major music festival cited as a future venue for interactive AI music experiences.

Jim Fan, lead of Nvidia's embodied AI research, outlines the transition from language-centric models to World Action Models (WAM) that simulate physical reality. He details a roadmap toward the 'Physical Turing Test' and autonomous factories by 2040, driven by video pre-training and human egocentric data scaling. ## [00:00] Introduction Host Sonya Huang introduces Jim Fan, who leads Nvidia's embodied autonomous research group. Fan reflects on his early days as an intern and the excitement surrounding the future of robotics. > *robots are just one of the most thrilling things that's going to happen.* > *[0, 12]* ## [00:30] DGX One Origin Story Jim Fan recounts the 2016 delivery of the first DGX-1 by Jensen Huang to Elon Musk and the OpenAI team. He highlights how this moment catalyzed the deep learning revolution that led to current AI breakthroughs. > *If you believe in deep learning, deep learning will believe in you.* > *[1, 26]* ## [01:42] The Great Parallel Fan proposes 'The Great Parallel,' applying the successful LLM scaling playbook to robotics. Instead of predicting the next token in a string, the goal is to predict the next physical world state through simulation and alignment. > *instead of simulating strings can we simulate next physical world state?* > *[2, 56]* ## [03:31] Robotics Endgame Setup The strategy for achieving the robotics end game is divided into two primary pillars: model strategy and data strategy. Fan notes that while LLMs are in their final 'boss fight,' robotics is just beginning its scaling journey. > *It boils down to two things, model strategy and data strategy.* > *[3, 32]* ## [03:39] Why VLA Falls Short Visual Language Action (VLA) models are criticized for being 'head-heavy' on language while lacking a fundamental grasp of physics and verbs. Fan argues they are better at encoding static knowledge than dynamic physical interaction. > *VLAs are great at encoding knowledge and nouns, but not so much at physics and verbs.* > *[4, 8]* ## [04:32] Video World Models Fan explains how video models like VEO3 learn internal physics—such as gravity and buoyancy—simply by predicting pixels at scale. These models act as simulators that can solve mazes and plan visual sequences internally. > *Physics emerge by predicting the next blob of pixels at scale.* > *[5, 15]* ## [06:09] DreamZero World Action Nvidia introduces 'Dreamer' and World Action Models (WAM), which jointly decode future world states and motor actions. This allows robots to perform zero-shot tasks by 'dreaming' the correct motion sequence before executing it. > *Dreamer jointly decodes the next world states and next actions.* > *[6, 29]* ## [07:46] Scaling Data Collection To overcome the physical limits of teleoperation, Fan discusses Universal Manipulation Interfaces (UMI) and exoskeletons like Dex-UMI. These tools allow humans to collect high-dexterity data directly without the robot being in the loop. > *we're able to break the curse of 24 hours per robot per day* > *[10, 6]* ## [11:06] EgoScale And Scaling Laws Fan introduces Ego-Exo, a policy trained on 21,000 hours of human egocentric video. This research uncovered a neural scaling law for dexterity, showing a mathematical relationship between pre-training volume and robot performance. > *we discovered this neural scaling law for dexterity.* > *[12, 39]* ## [15:39] DreamDojo And The Roadmap Fan outlines the roadmap to 2040, including the Physical Turing Test and 'lights-out' factories. He introduces Dream Dojo, a neural simulator that replaces classical physics engines with data-driven world models. > *I can say with 95% certainty that we'll get to the end of the end game... by 2040.* > *[19, 19]* ## Entities - **Jim Fan** (person): Lead of the embodied autonomous research group at Nvidia. - **Nvidia** (organization): The technology company developing the hardware and software for the robotics end game. - **Jensen Huang** (person): CEO of Nvidia, mentioned for delivering the first DGX-1 to OpenAI. - **OpenAI** (organization): The research lab that received the first DGX-1 for deep learning development. - **DGX-1** (product): The world's first deep learning supercomputer delivered in 2016. - **VEO3** (model): A video world model capable of simulating physics and visual planning. - **Dreamer** (model): A policy model that predicts future world states and actions simultaneously. - **Ego-Exo** (project): A robotics pre-training framework using large-scale human egocentric video data.

Andrej Karpathy explores the paradigm shift from traditional programming to Software 3.0, where LLMs act as programmable computers driven by context. He details the transition from 'vibe coding' to 'agentic engineering,' emphasizing that while AI handles execution, human taste and understanding remain the ultimate bottlenecks. ## [00:00] Introduction Stephanie Zhan introduces Andrej Karpathy, highlighting his foundational work at OpenAI and Tesla. She notes his unique ability to simplify complex AI shifts and introduces the concept of vibe coding. > *He has a rare gift of making the most complex technical shifts feel both accessible and inevitable. [00:22]* ## [00:44] Feeling Behind as a Coder Karpathy describes a turning point in December 2023 when agentic tools began producing perfect code without manual intervention. This shift led him to adopt vibe coding, trusting the AI to handle complex workflows autonomously. > *I just start to notice that with the latest models the chunks just came out fine. [01:29]* ## [02:28] Software 3.0 Explained Karpathy defines Software 3.0 as a paradigm where the LLM acts as a programmable computer and the context window serves as the primary programming lever. This follows Software 1.0's manual rules and Software 2.0's data-driven weight training. > *Software 3.0 is kind of about your programming now turns to prompting and what's in the context window is your lever. [03:20]* ## [03:44] Agents as the Installer Using the installation of OpenClaw as an example, Karpathy explains how agents replace rigid bash scripts with intelligent, environment-aware execution. This approach allows the AI to debug and adapt to specific system requirements autonomously. > *The agent has its own intelligence that it packages up and then it kind of like follows the instructions. [04:29]* ## [04:49] Menu Gen vs Raw Prompts Karpathy contrasts his custom-coded MenuGen app with raw prompts to models like Gemini, concluding that many traditional software layers are now redundant. He emphasizes that AI can now perform general information processing that was previously impossible with structured code. > *The software 3.0 paradigm is a lot more kind of raw. It just your neural network is doing more and more of the work. [06:11]* ## [07:37] What’s Obvious by 2026 Looking toward 2026, Karpathy envisions neural computers that process raw video and audio directly. These systems would use diffusion models to generate dynamic user interfaces, potentially making traditional UI code obsolete. > *You could imagine completely neural computers... a device that takes raw videos or audio into basically what's a neural net. [08:22]* ## [09:41] Verifiability and Jagged Skills AI models develop 'jagged' capabilities, peaking in verifiable domains like math and code due to reinforcement learning rewards. Karpathy notes the paradox where a model can refactor a massive codebase yet fail simple logic. > *state-of-the-art models today will tell you to walk [to a car wash] because it's so close... This is insane. [11:36]* ## [13:39] Founder Advice and Automation Model performance is heavily dictated by the specific data distributions chosen by frontier labs. Karpathy advises founders to explore the 'circuits' of these models to understand their strengths or use fine-tuning to fill gaps. > *we are slightly at the mercy of whatever the labs are doing, whatever they happen to put into the mix. [12:57]* ## [15:46] From Vibe Coding to Agent Engineering While 'vibe coding' raises the accessibility floor, 'agentic engineering' focuses on maintaining professional quality. This discipline involves coordinating powerful but stochastic agents to accelerate development without sacrificing the engineering bar. > *agentic engineering is about preserving the quality bar of what existed before in professional software. [16:07]* ## [25:17] Agents Everywhere and Learning Karpathy advocates for agent-native infrastructure, expressing frustration with human-centric documentation. He argues that while thinking can be outsourced to AI, human understanding remains a critical bottleneck for directing agents. > *You can outsource your thinking, but you can't outsource your understanding. [28:10]* ## Entities - **Andrej Karpathy** (person): AI researcher and former Director of AI at Tesla and founding member of OpenAI. - **Stephanie Zhan** (person): Partner at Sequoia Capital and host of the discussion. - **Software 3.0** (concept): A paradigm where LLMs act as programmable computers via prompting and context. - **Agentic Engineering** (concept): The professional discipline of coordinating AI agents to maintain software quality. - **MenuGen** (project): An app Karpathy built to OCR and visualize restaurant menus, used as a case study. - **OpenAI** (organization): AI research company co-founded by Karpathy. - **Gemini** (ai-model): Google's LLM used in Karpathy's software comparison. - **Vercel** (organization): A cloud platform used by Karpathy to deploy projects.