Sep 10, 2026 · 1h 22m · news
Jensen Huang Declares AGI Has Arrived | Tesla Launches Cybercabs | Index Pulls Out of Town
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this 20VC roundtable, Harry Stebbings, Rory O'Driscoll, and Jason Lemkin dissect the rapidly evolving AI landscape, analyzing frontier model economics, rogue autonomous agents, venture governance conflicts, and emerging tech liquidity pathways.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 12% of the talking time here. How this is scored →
Harry as informed peer 4.7 Guest teaching 4.5 Guest disagreement 3.4 Harry pushing back 2.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 11:43 Rory rejects AGI discourse as bullshit
Rory aggressively dismisses Jensen Huang's AGI declaration as a meaningless term, demanding people stop twisting themselves in pretzels and focus solely on coding LLMs.
Hardest push from Harry ▶ 14:09 Harry pushes back on legal AI labor compressionHarry refuses Rory's contention that law won't see the same AI take-rate as coding, directly countering that engineers were not eliminated either, merely redistributed.
Biggest teaching moment ▶ 48:21 Rory schools Harry on VC conflictsWhen Harry expresses surprise over Index passing on Town due to Instinct, Rory educates him on the vast operational difference between early-stage fiduciary board roles and late-stage passive investing.
Harry holds his own ▶ 1:13:40 Harry details Bret Taylor's market expansion pressureHarry demonstrates keen enterprise SaaS market insight by showing how Sierra's aggressive narrative expansion forces competitors like Wonderful to pivot from CX to full operating systems.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| AI Assistants, Instinct, and Rule-Breaking Startups | 6 | 3 | 4 | 4 | Harry makes sharp observations about form factor and WhatsApp distribution while pushing back on Jason regarding Instinct's pricing and Lovable parallels. Jason and Rory engage collaboratively on rule-breaking startup dynamics and whether large incumbents like Meta will simply clone these features. | |
| Jensen Huang Declares AGI and AI in Law vs. Coding | 6 | 7 | 5 | 5 | Harry argues that legal AI can rival coding AI in market size based on observing his girlfriend use Legora, but Rory firmly schools him on the economics, verifiability, and labor take-rates. Harry pushes back with 'I disagree', though Rory's detailed labor breakdown dominates the exchange. | |
| Model Fatigue, Benchmark Skepticism, and Fable 5.1 | 4 | 5 | 3 | 2 | Jason shares an extensive personal case study on using Fable 5.1 for debugging complex logic, contrasting it with performative benchmarks. Harry mostly listens and facilitates, stepping in at the end with Jacob Pachocki's alignment statement. | |
| AI Alignment, Cyber Threats, and Goal-Seeking Agents | 4 | 4 | 3 | 2 | Harry prompts the panel to discuss alignment, Pachocki's safety letter, and OpenAI's DSE wiki incident. Rory and Jason drive the conversation exploring the inevitable nature of goal-seeking agents finding cracks in guardrails. | |
| Tesla Cybercab, Waymo, and Travis Kalanick's Autonomy Bet | 5 | 3 | 3 | 3 | Rory tempers Harry's optimistic framing of Cybercab by noting it was an underwhelming 50-car trial in Austin. Harry injects London autonomous testing examples (Wave) to defend Uber's strategic timing in re-entering autonomy. | |
| Venture Conflicts: Index Pulls Out of Town Over Instinct | 4 | 6 | 5 | 3 | Harry finds Index backing out of Town due to Instinct surprising given platform-level investing. Rory directly corrects this framing, explaining the stark governance and signaling differences between early-stage boards and passive late-stage positions. | |
| Anthropic Walks from Descartes and the Toll of Broken M&A | 5 | 5 | 4 | 3 | Harry frames Anthropic walking from Descartes as potentially bad faith, which Rory quickly dismisses, explaining LOI versus definitive agreement due diligence realities. Jason elaborates on failed VC leak strategies damaging company value. | |
| Oura Ring IPO and Robinhood Entering Underwriting | 5 | 3 | 2 | 2 | Harry playfully notes Rory's mood about chosen topics before bringing up Oura's IPO and Robinhood's underwriting role. Rory and Jason explain the retail distribution upside and the desperate need for tech IPOs. | |
| Wonderful's $5B Valuation, Massive Secondary, and Strategy Pivots | 5 | 4 | 3 | 3 | Harry expresses shock at Wonderful's $170M secondary within two years, sparking an in-depth conversation on aggressive deal structures used by growth funds to win rounds. Harry correctly identifies Bret Taylor and Sierra forcing everyone to claim they are the full operating system. | |
| Thinking Machines, Poolside's Nvidia Sale, and Neolab Survival | 5 | 4 | 3 | 2 | Harry introduces Thinking Machines' $40B valuation and outlines the Neolab funding crunch. Rory and Jason detail the contrast between Poolside's forced asset sale to Nvidia and Thinking Machines' ability to raise corporate capital. | |
| Anthropic S-1 Filing Rumors and Public Index Exposure | 3 | 5 | 2 | 2 | In a brief closing banter, Rory mentions Twitter rumors of Anthropic's S-1 and gives Harry a mini-lesson on index exposure, showing that passive QQQ/SPY holders will own the stock automatically. |
Statements from this episode (35)
Lemkin: GrokBot violates Google's terms of service to deliver answers
“GrokBot uses Google, which is not allowed, which is prohibited by the term of service to Google things and then give you answers.”
O'Driscoll: No scaled business is built on scraping LinkedIn or Google
“The LinkedIn scraping in the end, no business at scale ever gets built on that, and no business at scale ever gets built on breaking Google's terms of service.”
Lemkin: Public company CEOs say legal teams block them from competing
“I I've talked with multiple public company CEOs who say they're hamstrung, their hands are tied behind their backs. They can't compete with startups, not because they can't do it. Okay. Rory O'Driscoll: That's true. Jason Lemkin: But because their legal teams …”
Lemkin: Gorgias' WhatsApp AI agent reached double-digit usage in weeks
“I invested years ago in a company called Gorgias, which is a little over a hundred million, which used to be So e-commerce support. Now it's an AICX, right? And they launched their agent in WhatsApp that anybody can uses and text, and it's already double digit…”
Lemkin: Passing on Instinct at a $2.5B valuation
“The last round was two and a half, Jason. Jason Lemkin: I know. That's why we're going to have to pass on the round.”
Stebbings: Buy Meta due to its distribution moat for AI agents
“My take is actually it's why you buy Meta today, because you've got the most clear, unwavering PMF for this product. And Zuck is the one who owns the core distribution channel, and Zuck has been working on this product.”
O'Driscoll: Coding is a half-trillion dollar industry
“The only thing that mattered for the last two years is LLMs do code, and code is a half a trillion dollar industry.”
O'Driscoll predicts anything reducible to code will be automated by LLMs
“Anything that can be reduced to code will be done by it.”
Stebbings: AI in legal services could be a half-trillion dollar market
“If coding is a half a trillion dollar market, and you have two companies like Harvey and Legora, I don't see why there's not a half trillion dollar market in law.”
Lemkin discloses his team runs Replit 20 hours a day
“Even for me, and again, I know folks sometimes mock me, but the biggest change for our little tiny team of two and a half humans is we run Replit between me and Amelia. We run it 20 hours a day now.”
O'Driscoll predicts AI captures 10-15% of legal spend versus 30-50% in coding
“Your type A lawyers will use it 24 seven, but my guess is 10, 15% of total spend goes to AI, whereas in law and in coding, you can argue for 3040, 50%, right?”
O'Driscoll: Coding is inherently more verifiable than legal work
“Coding is inherently more verifiable. Some parts of it are mathematically verifiable. Some parts you can just run on the machine and confirm.”
Lemkin: CEO AI benchmarks on X are essentially worthless
“And you look on X and all the CEOs are sharing their benchmarks, which are essentially worthless, right? There's no cost or time in them and it's all performative AI, like vibe coding your own CRM.”
Lemkin: Fable 5.1 is a step function comparable to an S-tier CTO
“Fable five for me was that for just for me was that step function where all of a sudden we could solve big problems together the way you'd like to with like your best CTO, like if you've ever worked with a five out of five or like an S tier CTO where you could…”
O'Driscoll: Economic value will replace AGI questions and benchmarks
“We'll know what works because people will use it because people are rational and economic actors, right? And all these questions on AGI and benchmarks will be replaced by the question, is this model the one that makes me the most, you know, that generates the …”
O'Driscoll: Advanced AI models have had a massive impact on cyber risk
“I mean, I think to be fair, unlike some of the other PDOOM stuff, right, there's real evidence that the impact of these models on cyber risk has been massive.”
Lemkin: Autonomous AI agents hit a 'Dunbar number' where rules inevitably conflict
“If you have too many rules, they always conflict. It's almost unsolvable. There's some number, I don't know, there's some number of Dunbar, there's a Dunbar number for rules where you get out to 40, 50, 60, 70 gates on a process. Poor instinct and GrokBot can'…”
O'Driscoll: Tesla is the only other credible competitor to Waymo
“They're the only other competitor to Waymo.
With credibility.”
O'Driscoll: Waymo brings in hundreds of millions, not billions, in revenue
“There are, you know, hundreds of millions of dollars in revenue, but not billions.”
Lemkin: I sold my car and rely solely on Waymo and Uber Black
“In the Bay Area, I got rid of my car, so I only do Waymo and autonomous driving.
Yeah.
I don't drive.
I'm done with it.
if there is an issue, I take a Uber black, but I don't have a car anymore in the bear.
I just don't have one.”
Lemkin: Autonomous transit will replace urban car ownership within 10 years
“So I think, you know, it's probably another 10 years where anyone with a brain is, is gonna have this be their primary
Mode of ownership if you're not in the country or you don't have niche use cases.”
Stebbings: Index dropped Town deal after portfolio company Instinct objected
“There was a conflict in venture that has now prevented a deal.
We've spoken about instinct being the AI assistant that's raised from index and benchmark.
Well, there's another AI assistant called town, which we just had on the show and index.
Were going to …”
O'Driscoll: Holding competing late-stage AI stakes creates no real conflict
“It totally makes sense to be In OpenAI and Antropic at two hundred billion pre each time, you get limited information rights, retroactive information only, and you know, you're just on the cap table, and it's no different than being in two public companies.”
O'Driscoll: No one could serve on both OpenAI and Anthropic boards
“I don't think anyone could be on the board of Antropic and also on the board of OpenAI.”
O'Driscoll: Leaking acquisition talks leaves companies shop-spoiled if deals collapse
“That way, once it leaks, even if you leak it to, as the company being acquired to drum up competitive bid, the problem is you've set yourself up for this thing whereby if subsequently the deal doesn't come together, you look a bit, you know, shop spoiled, for …”
Lemkin: Video diffusion is not a top-three use case for Anthropic
“It's not a core. It's not one of their top three use cases for the LM is video diffusion.”
O'Driscoll: Public foundation model companies will acquire neo-labs for TAM expansion
“Some of them might, will be great investments, because I do think the foundation model companies, once the public will be acquirers of some of this stuff for time expansion, but they won't all be great investments.”
Stebbings: Robinhood makes underwriting debut listed 18th on Oura's IPO
“One is this Robin Hood's first role as an underwriter. They're listed 18th and last.”
Lemkin: Oura will have a pretty successful IPO
“I think it's going to be a pretty successful IPO. It's the kind of thing people are going to want to buy. They understand it and the growth. It's not 20% growth. This is not 18% growth and there's downside. There's competition. It's confusing, but the subscrip…”
O'Driscoll: Large secondary sales serve as essential recruiting tools for enterprise deployment
“From a recruiting perspective, it allows you to say to the next hundred people, come to work with us. Yes, you'll get stuck on a five month deployment on a bank in Holland or a, you know, a electrical company in Germany. It'll be boring as shit, but in return,…”
Lemkin: VCs will increasingly offer bad deal structures to win competitive rounds
“We will see deal structures that are objectively bad for the company done more and more often to win deals, whatever it takes.”
O'Driscoll: Sierra must become a complete customer ecosystem to justify massive valuation
“The only way I can justify 15 or twenty billion in market cap for Sierra is not if I build a slightly better next generation service cloud, if I am the entire customer ecosystem for your entire business.”
Lemkin: Funding will end for many AI neolabs
“It is it is a reminder that the music will end for a lot of the neolabs.”
O'Driscoll alleges Poolside effectively cashed out via Nvidia product license
“A few weeks ago, Poolside effectively didn't, they would deny that they sold, but they sold a license for their product to NVIDIA, which was an open weight enterprise US focused model, and they, you know, the company still exists, but they cashed out quite a l…”
O'Driscoll: Corporate America will drive demand for US open-weight AI
“I mean, fundamentally, you do believe that there is going to be strong demand from corporate America for a open-weight US-based model with the infrastructure to train that model, right? And I think Thinking Machines is in a good position to meet that demand no…”