Apr 15, 2024 · 53m · news
Sam Altman & Brad Lightcap: Which Companies Will Be Steamrolled by OpenAI? | E1140 · 20VC with Harry Stebbings
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 joint interview on the 20VC podcast, OpenAI CEO Sam Altman and COO Brad Lightcap discuss their leadership partnership, the predictive scaling laws of AI, and company scaling strategies. They share insights into technical roadblocks, warn startups against building static software wrappers, and discuss the future of global intelligence abundance.
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 17.8% of the talking time here. How this is scored →
Harry as informed peer 2.5 Guest teaching 3.0 Guest disagreement 2.1 Harry pushing back 2.5
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 16:08 Sam dismisses marginal cost question
Sam forcefully rejects Harry's premise by calling the question about marginal cost versus revenue 'the most boring question I could imagine.'
Hardest push from Harry ▶ 26:03 Harry challenges iterative deployment feasibilityHarry directly challenges Sam's release philosophy by pointing out that companies like Meta and Google faced severe stock and reputation hits when deploying imperfect models.
Biggest teaching moment ▶ 41:58 Sam redefines CEO communicationSam reframes Harry's assumption that founder communication means media polish or charisma, correcting it to functional operational skills like explaining vision, recruiting, and selling.
Harry holds his own ▶ 41:58 Harry pushes back using VC deal experienceHarry leverages his experience as a seed investor to challenge Sam's criteria, arguing that elite engineering founders often lack communication skills early on.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| The Origins of OpenAI and Sam Altman's Early Conviction | 1 | 2 | 1 | 1 | Harry sets up friendly context questions about the founding conviction behind OpenAI. Sam explains early beliefs in deep learning and scaling laws in an agreeable tone. | |
| The Origin and History of the Altman-Lightcap Partnership | 1 | 2 | 1 | 1 | Harry asks how the partnership originated. Brad tells a humorous story about failing to recruit 25 candidates for CFO before taking the role himself. | |
| Leadership Roles and Complementary Strengths | 1 | 1 | 1 | 1 | The host asks each guest to highlight the other's hidden strengths. Sam commends Brad's adaptability while Brad highlights Sam's long-term orientation and laser focus. | |
| OpenAI's Core Priorities and the Compute Constraint | 2 | 2 | 2 | 2 | Harry inquires about bottlenecks to OpenAI's velocity. Sam points to compute supply and jokingly dodges revealing specific compute strategy on camera. | |
| Decision-Making, Delegation, and the Operator's Reality | 2 | 3 | 2 | 3 | Harry asks whether companies are defined by macro decisions or daily micro decisions. Sam explains operator reality, admitting candidly that he is not naturally an operator. | |
| Compute Economics, Open Source, and the Abundance of Intelligence | 3 | 5 | 6 | 3 | Sam bluntly rejects Harry's question about marginal costs as the most boring question imaginable, explaining why falling compute costs render the premise trivial. | |
| Model Commoditization and Base Model Differentiation | 3 | 6 | 4 | 2 | Harry asks about model commoditization and startup viability. Sam explicitly warns that startups building on the assumption that base models will not improve will get steamrolled. | |
| The Iterative Deployment Philosophy | 4 | 4 | 2 | 5 | Harry challenges Sam's iterative deployment strategy by citing major public backlash suffered by Meta and Google for releasing imperfect models. | |
| Scientific Progress as the Ultimate High-Order Bit | 2 | 3 | 2 | 2 | Sam discusses AI driving scientific advancement while noting that current models remain too weak, while Brad outlines the diverse real-world adoption of ChatGPT. | |
| Filtering Talent: Mission-Driven Hires vs. Mercenaries | 3 | 3 | 2 | 4 | Harry presses Sam for specific founder lessons, leading Sam to highlight insights learned from Brian Chesky and the Collison brothers. | |
| Enterprise ROI vs. Unlocking the Supply of Time | 4 | 4 | 3 | 3 | Brad explains enterprise adoption misconceptions regarding ROI. Sam pokes lighthearted fun at enterprise jargon when Brad mentions quantifiable business processes. | |
| Cultivating Harmony Between Research and Sales | 3 | 3 | 2 | 3 | Harry asks how OpenAI balances research culture with sales expansion. Sam explains alignment with Brad and reflects on why unprecedented growth offers few repeatable lessons. | |
| Evaluating Outlier Founders: The Power of Clear Communication | 5 | 5 | 3 | 5 | Harry draws on his venture capital experience to dispute Sam's emphasis on communication skills for technical founders. Sam redefines communication as functional execution rather than media charisma. | |
| Experience vs. Raw Hustle in OpenAI's Recruitment Strategy | 4 | 3 | 2 | 3 | Harry asks if OpenAI skews older and favors experience over hustle. Sam and Brad explain that novel domain problems require open-minded talent alongside deep research experience. | |
| Introduction to the Quick Fire Round | 2 | 2 | 1 | 2 | Harry conducts a quick-fire round covering 12-month goals, enterprise speed, global concerns, and model scaling predictability. | |
| Personal Sacrifices and Running Out of Time | 1 | 1 | 1 | 2 | Harry asks personal questions regarding sacrifices and relationships. Sam shares a candid reflection on running out of time for personal life while remaining deeply happy. | |
| The Future State of the World and Technological Abundance | 1 | 2 | 1 | 1 | The interview concludes with optimistic projections about future technological abundance eliminating illness and scarcity. |
Statements from this episode (35)
Altman: 95% of AI startups should bet on models continuing to improve
“It would seem to me that 95% of the world should be betting on the latter category, but a lot of the startups have been built in the former category.”
Altman: OpenAI will steamroll startups building static software wrappers
“When we just do our fundamental job because we, like, have a mission, we're going to steamroll you.”
Altman: OpenAI was founded on the insight that deep learning improves with scale
“Deep learning seemed to actually legitimately be working. And two, it got better with scale. We didn't know how predictably it got better with scale at the time, but it was clear that like bigger was better.”
Lightcap: 25 candidates rejected the OpenAI CFO position before he took it
“I asked probably 25 people if they would want to be CFO of OpenAI, which at the time was just a small kind of non, sleepy non-profit, and I went over 25, and honest to God, the reason I'm here is because I was so embarrassed to come back Like, over 25 that I s…”
Altman: OpenAI had no appreciable business until very recently
“We didn't have a business at all, or we didn't have an appreciable business until very recently.”
Altman: OpenAI's rapid enterprise go-to-market success surprised him
“I think if you had said a year ago, we're going to be like a great organization. Oh, not yet a great organization. We're going to even be a very good organization at doing You know, an enterprise go-to-market function, I would have said, ah, very low chance th…”
Altman: OpenAI will surprise the world by solving its compute constraints
“By treating that as a whole system problem I am optimistic we will really surprise the world on the upside.”
Altman: Venture investing comes down to one or two decisions a decade
“One of the things that I loved about being an investor was that job is really a job about one or two decisions a year, or maybe one or two decisions a decade.”
Altman: Venture capital offers a great quality-of-life trade-off
“For a quality of life job, it's a great, great trade-off.”
Altman: AGI will be the most important thing I ever touch
“With no false humility, I'm just not an operator by nature. I'm happy to do it, because I, like, really love OpenAI, and I think AGI will be the most important thing I ever touch, but this is not my natural fit.”
Altman: OpenAI will drive the cost of high-quality intelligence near zero
“I think we can drive the cost of a very high quality of intelligence to very near zero, and that will just be phenomenal for most things in the world. Not everything, there will be some negatives, but I think the cost of intelligence is about to get really, re…”
Lightcap: Current AI models are honestly not that good
“I think right now expectations are extremely high. Reality is still pretty bad, honestly. These models are not that good.”
Lightcap: AI expectations will dip before rapid model improvements outpace them
“I think very quickly expectations will start to come down as people come into contact with today's models, but then very quickly also these models will get really, really good, and you'll see this inversion of expectations reality where all of a sudden then ex…”
Altman: The frontier AI market will consolidate into roughly a dozen providers
“I think eventually it will shake out. There will be a small number of Providers, just a relatively small number, you know, dozen, something like that, doing models at big scale, and it'll be extremely complex, extremely expensive”
Altman: Long-term AI differentiation will come from personalization, not base models
“I don't think the base model, like that's just, you know, intelligence is just like some emergent property of matter or something. The long-term differentiation will be the model that's most personalized to you, that has your whole life context, that plugs int…”
Stebbings: 20VC is not investing in any AI application layer startups
“For me as an investor, I see so many AI companies, and I'm not investing in any application layer of AI companies.”
Altman: Deploying ChatGPT was one of OpenAI's most important decisions ever
“In some sense, one of the most important decisions we ever made was this one. And that includes things like deploying ChatGPT into the world and getting the world to take advanced AI seriously, which we tried to talk about for a long time and didn't really wor…”
Altman: AI will meaningfully increase the rate of scientific progress
“And if AI can help people meaningfully increase the rate of scientific progress, which I believe it will, I think that will be a triumph.”
Altman: Raw model intelligence is the primary bottleneck to AI accelerating science
“I think the models are just not smart enough, which sounds like a annoying, low information kind of cop-out answer, but I think it's like deeply fundamentally true. Like the models just aren't smart enough. You fix that one thing, all these other things get be…”
Altman: GPT-6 will be a general-purpose tool for scientific research
“And with GPT-VI, I think people will say, hey, this is like helping me as a general purpose tool in all these ways.”
Altman: Airbnb's Brian Chesky heavily advised OpenAI on product and messaging
“Chesky has been incredibly hands-on and helpful to me over the last year and a half and is really good at a lot of things that I'm not good at, and I've had to, like, come up to speed quickly on how to think about how we talk about our products how to think ab…”
Lightcap: Enterprises criminally underrate the ROI of granting broad AI access
“I think people though criminally underrate how important it is actually and how much like return you really get on just giving people access to the technology. And that there's this kind of, because you can'queries quite quantify exactly how it works, but like…”
Lightcap: Many enterprises mistakenly assume GPT-4 is the peak of AI
“A lot of companies think it's static, so a lot of companies think GPT-IV is the best the models will ever get. That's understandable. Every technology they've ever had to adopt has been relatively static.”
Altman: Entrepreneurs learn far more from moderate success than from failure
“I think you usually don't learn that much from failure. You learn more from success. But I think you also don't learn that much from, like, extreme break all the rules, unrepeatable success either.”
Altman: ChatGPT's explosive growth offers no actionable growth lessons for startups
“And what we had with ChatGPT, I would be hesitant to say I've learned anything at all about growth. Like have a once in a generation technological revolution. It's not really like actionable advice.”
Lightcap: Company-changing ideas usually come from less experienced team members
“Really that, that, like, company changing ideas actually, by and large, come from places that are not, not those hires.”
Altman: OpenAI's leadership and technical teams skew older than typical startups
“Our leadership team is probably More like thirties and forties than the twenties and thirties you would see at other startups. And I think our technical people skew like slightly older I don't have numbers, but you know, maybe I would guess that like the avera…”
Altman: Securing compute and supply chain is OpenAI's biggest five-year challenge
“Doing the best research and the best productization of, like, the best innovation on that stuff over the next 12 months, and was it five years for the second there? Sufficient, like, supply chain and compute.”
Lightcap: Enterprise AI spending is now drawn from real corporate budgets
“Well, we have real budgets and that'll help.”
Altman: Global geopolitical instability is the highest he has ever seen
“The whole thing just feels like way more on the whole situation of the world, the geopolitical thing, the sort of socioeconomic stuff, politics, it feels more unstable to me than it has felt since I've been paying attention. And there's no, like, one thing I w…”
Lightcap: AI model scaling laws have held predictably consistent for six years
“I think it's how consistently the scaling of models has worked. It still breaks my brain. Like I don't, maybe I've been, if we've, I've watched the same trend line for six years now, but I still find it incredible that you can make these models bigger and they…”
Lightcap: OpenAI was surprised by AI's rapid impact on creative industries
“It caught us somewhat by surprise how important the technology is and is going to be in creative industries, for example relative to more knowledge based industries or relative to even more industrial industries.”
Lightcap: OpenAI originally focused on robotics and gaming before pivoting entirely
“We were doing robotics really early on. And so I was preparing for a world where we were working with robotics companies and building robots and working with gaming companies and building agents. And we've gone completely the other way.”
Altman: Leading OpenAI has required sacrificing his entire personal life
“I'm okay with this trade for now, because I know it's not a forever thing, but I have basically, like, run out of time for real life. I don't really get to hang out with friends that much. I don't get to, like, do the normal, like, life stuff.”
Shorts cut from this episode
▶ Top things to look for when investing 💰 · 20VC with Harry S (@41:11)
▶ Who are the best Founders in the world? 🌎 · 20VC with Harry (@32:32)
▶ #1 skill Founders need 🙌 · 20VC with Harry Stebbings (@42:03)
▶ Why Sam Altman is world class 🏆 · 20VC with Harry Stebbings (@9:02)
▶ How I became CFO for OpenAI ✍️ · 20VC with Harry Stebbings (@5:03)
▶ OpenAI CEO: How to make decisions 💡 · 20VC with Harry Stebb (@14:00)