Sep 29, 2022 · 57m · how-i-built-this
HIBT Lab! OpenAI: Sam Altman
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of How I Built This Lab, Guy Raz interviews OpenAI co-founder and CEO Sam Altman about the genesis, rapid evolution, and societal implications of Artificial General Intelligence (AGI). Altman discusses his entrepreneurial career, OpenAI's novel organizational structure, and the technical and ethical principles required to guide powerful AI systems toward the collective benefit of humanity.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Guy holds 38.2% of the talking time here. How this is scored →
Guy as informed peer 4.1 Guest teaching 2.1 Guest disagreement 1.1 Guy pushing back 1.3
speaking balance: gold is Guy, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 53:19 Altman rejects evolutionary replacement framing
Altman directly dismisses Raz's comparison of AGI to evolutionary species replacement, arguing human intentionality makes the process incomparable to random biological selection.
Hardest push from Guy ▶ 33:27 Raz challenges capped-profit pivot and Microsoft dealRaz directly questions whether accepting Microsoft investment and granting exclusive software licenses undermines OpenAI's founding promise of open research.
Biggest teaching moment ▶ 26:28 Altman deconstructs the robotics fallacyAltman dismantles the impression that solving a Rubik's cube was a breakthrough, educating the listener that physical robotics work stalled due to inadequate simulation environments.
Guy holds their own ▶ 37:11 Raz demonstrates GPT-3 factual hallucinationsRaz showcases hands-on testing of OpenAI's model by reading two completely fabricated biographical summaries generated by GPT-3 about Altman himself.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Guy as informed peer | Guest teaching | Guest disagreement | Guy pushing back | Why |
|---|---|---|---|---|---|---|
| Sam Altman's Early Fascination with Computing | 3 | 1 | 0 | 0 | Guy Raz sets a relaxed biographical tone, prompting Sam Altman to recall his childhood fascination with early computing and ham radio. Altman enthusiastically reflects on the magic of connectivity without any conflict. | |
| Stanford CS and Dropping Out for Loopt | 3 | 1 | 1 | 0 | Raz asks about Altman's decision to drop out of Stanford for Loopt and adds context about mid-2000s mobile tech. Altman notes that Loopt's core assumption about spontaneous human mobility turned out largely wrong. | |
| The Y Combinator Experience and Loopt's Acquisition | 3 | 1 | 0 | 0 | Altman recounts entering Y Combinator's inaugural batch and learning business development under Paul Graham's mentorship. Raz facilitates the narrative smoothly as Altman reflects on angel investing returns versus startup exits. | |
| Leading Y Combinator and Championing Hard Tech | 4 | 1 | 1 | 0 | Raz notes Y Combinator's historical focus on software while Altman explains his push into deep tech and hard science during his presidency. Altman describes his initial reluctance toward becoming an investor before embracing the platform. | |
| The Genesis of OpenAI and Vision for AGI | 4 | 3 | 1 | 1 | Altman details the origins of OpenAI after the 2012 deep learning breakthrough. When Raz asks about ImageNet searching text for images, Altman gently clarifies that ImageNet was about classifying given images into categories. | |
| OpenAI's Nonprofit Foundation and Mission | 4 | 2 | 0 | 0 | Raz asks why OpenAI was structured as a nonprofit given the immense capital requirements. Altman explains the moral rationale for broad public benefit and notes that early AGI research was initially dismissed by mainstream researchers. | |
| Envisioning Everyday and Global Applications of AGI | 3 | 2 | 0 | 0 | Altman outlines his vision of AGI as a companion assistant and a tool for macroscopic scientific breakthroughs like cancer research. Raz invites him to paint a concrete picture of future workflows. | |
| Reinforcement Learning in Gaming and Robotics | 4 | 3 | 1 | 1 | Raz asks how solving a Rubik's cube with Dactyl advanced OpenAI's mission. Altman candidly admits the robotics initiative turned out to be a strategic mistake due to poor simulation environments and hardware limits. | |
| Managing Technological Uncertainty vs. Fusion Energy | 6 | 3 | 2 | 5 | Raz presses Altman on transitioning OpenAI into a capped-profit entity and granting exclusive licenses to Microsoft, asking if this compromises their open-source mission. Altman explains the massive compute expenses and details safety governance mechanisms. | |
| Language Models, Hallucinations, and DALL-E | 6 | 2 | 1 | 2 | Raz reveals that he tested GPT-3 by prompting it to write Altman's biography, finding plausible but totally fabricated biographical details. Altman agrees that hallucination and lack of verification remain large language models' biggest challenge. | |
| Mitigating Risks and Aligning AI Incentives | 5 | 2 | 2 | 3 | Raz draws a comparison to early social media optimism that soured and asks whether catastrophic misuse by bad actors is inevitable. Altman rejects defeatism, insisting on cautious rollouts and aligned institutional incentives. | |
| Ubiquitous Intelligence and Creative Job Automation | 4 | 3 | 1 | 1 | Raz and Altman discuss how AI unexpectedly automated creative and white-collar tasks before blue-collar labor. Altman notes that expert consensus five years earlier got the technological roadmap completely backwards. | |
| Labor Redefinition and the Future of Human Purpose | 4 | 2 | 2 | 1 | Raz asks about societal instability once AI displaces workers, and Altman pushes back against pearl-clutching, arguing traditional work should eventually become optional while humans will continuously invent new pursuits. | |
| Evolutionary Agency and Ethical Responsibility | 5 | 3 | 3 | 4 | Raz asks whether AGI represents a new evolutionary species that will displace Homo sapiens and asks Altman how we know he won't become a regretful Oppenheimer. Altman rejects the evolutionary comparison and acknowledges keeping an Oppenheimer biography on his desk. |
Statements from this episode (25)
Altman: Location apps struggled because adults have far less spontaneity
“Even as the world has gotten very mobile, most people are sort of still at home or work all the time, and there's just, like, far less spontaneity than it felt like to, like, a, you know, twenty-year-old or whatever I was at the time.”
Altman: Angel investing paid orders of magnitude more than founding Loopt
“I made like orders of magnitude more money from angel investments that I made. That I spent no time on the thing I poured my life into.”
Altman: Becoming a VC felt like an admission of defeat
“I didn't, like, have a ton of respect for the career path. And so there was, like, some sense in which it was an admission of defeat about, you know, I just, I can't run a company, so I'm gonna go do the easy job, or I'm gonna go do the retirement job or somet…”
Altman: Silicon Valley was not geared toward capital-intensive science
“There's a thing about Silicon Valley, which is it's, it had not been that well geared towards the time and capital intensive science projects.”
Altman: Deep neural networks showed performance scaling with compute in 2012
“And then in 2012 deep neural networks started to work. And not only did they start to work, it appeared that the more compute you threw at them, the better they got.”
Altman: AGI will likely be the most important invention in human history
“And I think if we're able to accomplish that, it will be perhaps the most important invention in human history, and it will be the culmination of this phenomenal amount of effort all the way down the stack and like the collective knowledge, accomplishment, mor…”
Altman: In 2016 researchers believed AGI was a century away
“At the time, People thought we were crazy. Like, good researchers say you've totally discredited yourself by talking about AGI. Most other people were like, this is, you know, a hundred years away or more.”
Altman: AGI is an exception to capitalism and should be government-led
“Obviously I think capitalism is great. But I think AGI is sort of an exception to that. In a well-run society, I think even if we rewound the US, you know, 50 plus years, this would absolutely happen by the government.”
Altman: AGI will eventually be capable of finding a cure for cancer
“Eventually, now this will require more compute, but if we society at allocate to compute to that, you can say, like, could you go off and find a cure for cancer? And it can spend a lot of its compute cycles doing amazing scientific progress like that.”
Altman: Everyone will eventually have an AGI companion assisting them all day
“But I imagine that we get to a world where each of us has sort of like a, what feels like an AGI companion that we are talking to all day that is helping us be the best version of ourselves, learn, be efficient, be happier, and we'll all experience that way.”
Altman: OpenAI paused robotics research because hardware and simulators were too limited
“Well, it turned out to be a mistake. I mentioned that Dota was a great environment. It turns out that robotics are a very hard environment. Robotics is hard, and it's not hard because the machine learning is hard. It's hard because the simulator is bad, the ro…”
Altman: OpenAI picks projects by following what works, not a grand plan
“The way we pick projects to work on is, like, not as exciting as everybody would hope. You know, I think there's this belief that you have a bunch of people, like, sitting in a room picking this, like, brilliant secret strategy. We just sort of, like, run our …”
Altman: Helion Has Clear Roadmap Certainty While OpenAI Operates in Uncertainty
“The other company I'm involved with is this nuclear fusion company called Helion. And there are a lot of similarities between the two companies in terms of like a very hard Scientific and engineering problem. But the biggest difference is like what to do, what…”
Altman: The government refused to fund OpenAI
“We did also just see if the government wanted to fund us. They definitely did not.”
Altman: OpenAI's board can wipe equity to zero for safety
“We also have something in our documents which says if we need to for safety, the board can just totally wipe out everyone's equity value to nothing.”
Altman: OpenAI charter permits shutting down to merge with leading rivals
“And we have something called the merge and assist clause, which says, similar to that, if another effort's ahead, and we want to avoid a race condition we can just shut down and merge with some other effort.”
Altman: Open-sourcing AI model weights is an irreversible one-way door
“If we just publish the weights of a model on the internet, and then we realize, like there's actually a safety issue here, we can't take that back. It's done. It's like a one-way door. We also cannot put any usage restrictions on it after we open source it.”
Altman: Models orders of magnitude above GPT-3 could enable massive disinformation
“Like, at some point there will come a time, one, two, three orders magnitude, who knows what, how much more powerful than GPT-III, where I think a language model Can really have a huge disinformation effect on the world.”
Altman: Powerful digital intelligence will exist within 10 years
“Sometimes these things take longer than you think, but let's say, like, in 10 years, I think there will be powerful digital intelligence in the world.”
Altman: OpenAI deploys technology slowly and cautiously, even if users get frustrated
“Right now, you know, we deploy our technology slowly and cautiously, like we're willing to piss users off to go slowly, but as the systems get much more powerful, the challenges become more and more unprecedented.”
Altman: AI will become a ubiquitous layer across all software
“And I think the same thing will happen for AI. There will be a layer of intelligence provided by us and others. It's just everywhere. And as the systems continually get smarter, everything you use will get smarter and smarter as well.”
Altman: AI is automating creative jobs before blue-collar labor
“You know, the strong consensus Five, 10 years ago, was that first of all the AI was going to come for the blue-collar jobs. Second of all would be the less sophisticated white-collar jobs. Third of all would be the very high cognitive load white-collar jobs li…”
Altman: Traditional work should be optional
“I think, like, traditional work, in the way we think of it, should be optional, and you should be able to do less of it.”
Altman: Human creativity and desire for status will not disappear with AI
“I don't know what the jobs of the future will look like, but I am confident that human creativity Desire for status. Desire to, like, do new things and to, like, accomplish. That's not gonna go anywhere.”
Altman: The best long-term future involves humans merging with AI
“And the most positive long-term futures that I can imagine involve some degree of a merge. It doesn't have to be the crazy sci-fi, like, you know, plug something into our brains or upload ourselves or whatever. But at a minimum, I think it needs to involve som…”