
50 segments available
Nat Friedman was the CEO of Github from 2018 to 2021. Before that, he started and sold two companies - Ximian and Xamarin. He is also the founder of AI Grant and California YIMBY. And most recently, he has created and funded the Vesuvius Challenge - a million dollar prize for reading an unopened Herculaneum scroll for the very first time. If we can uncover these scrolls, we may be able to recover lost gospels, forgotten epics, and missing works of Aristotle. We also discuss the future of open source and AI, running Github and building Copilot, and why EMH is a lie. Vesuvius Challenge: https://scrollprize.org/ Nat's Wisdom: https://nat.org/ Nat's Twitter: https://twitter.com/natfriedman 𝐄𝐏𝐈𝐒𝐎𝐃𝐄 𝐋𝐈𝐍𝐊𝐒 * Transcript: https://www.dwarkeshpatel.com/p/nat-friedman * Apple Podcasts: https://apple.co/3TCHFcY * Spotify: https://spoti.fi/42rce9v 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 00:00:00 - Vesuvius Challenge 00:30:00 - Finding points of leverage 00:37:39 - Open Source in AI 00:40:32 - Github Acquisition 00:50:18 - Copilot origin Story 01:11:47 - Nat.org 01:32:56 - Questions from Twitter
Nat Friedman introduces the Vesuvius Challenge, a groundbreaking initiative aimed at reading the ancient Herculaneum scrolls buried by the eruption of Mount Vesuvius in AD 79. He shares the historical significance of Herculaneum, the discovery of its buried treasures, and the challenges faced in accessing and reading these fragile scrolls, which could potentially double our knowledge of ancient texts.
"the Scroll prize, which is where we'll start this conversation. Do you want to tell the audience about what the Scroll prize is? We're calling it the Vesuvius challenge. It is just this crazy ..."
Friedman recounts the dramatic history of Herculaneum, detailing how the town was buried under volcanic ash and mud, preserving a unique library of papyrus scrolls. He explains the initial discoveries made in the 1700s, the looting that occurred, and the subsequent realization that these scrolls contained valuable texts from antiquity, despite their fragile condition.
"it was a very thick layer. Those towns were buried and forgotten for hundreds of years. No one even knew exactly where they were, until the 1700s. In 1750 a farm worker who was digging a well in t..."
Nat discusses the difficulties faced in attempting to unroll the carbonized scrolls without causing further damage. He highlights the historical attempts to read these scrolls, including the use of a machine developed by an Italian monk, and the unfortunate destruction of many scrolls due to misguided efforts. The segment emphasizes the fragility of these ancient texts and the need for innovative solutions.
"than 1% of all the writing from that period. It was an enormous discovery to find these hundreds of papyrus scrolls underground. Even if they were not in good condition but still present. On a fe..."
Friedman shares insights into modern technological advancements aimed at virtually unrolling the Herculaneum scrolls. He highlights the work of Professor Brent Seales and his team, who are using advanced imaging techniques and machine learning to decode the scrolls without physically opening them, marking a significant step forward in the field of archaeology.
"to physically unroll these things because so many were destroyed. In fact, some attempts to physically unroll the scrolls continued even into the 200s and they were destroyed. The current situati..."
In this segment, Nat outlines the potential impact of successfully reading the Herculaneum scrolls, which could double the existing corpus of ancient texts. He discusses the historical significance of the texts that may be contained within these scrolls, including lost works of literature and philosophy, and expresses his excitement about the possibilities that lie ahead.
"lost information from the past And I waited for the news that they had been decoded successfully. That was 2020 and then COVID hit, everybody got a little bit slowed down by that. Last year, I f..."
Friedman reflects on the extraordinary nature of the Vesuvius Challenge, likening it to a science fiction narrative. He shares anecdotes about his collaboration with Professor Seales and the excitement surrounding the potential discoveries that could emerge from the scrolls, emphasizing the historical and cultural significance of unlocking these ancient texts.
"that I've heard. So that would be significant. We don't know what's in there, we've got a few Philodemus texts, those are of some interest. But there could be lost epic poems, or God knows what. ..."
Nat provides context on the excavation of the Herculaneum villa, discussing the limited exploration that has occurred and the potential for undiscovered scrolls. He shares insights from historians about the possibility of a larger library containing significant literary and historical texts, and the romantic notion of what might still be hidden beneath the layers of volcanic material.
"he finally got on the phone with me and we got to zoom. And he's like, why are you interested in this? I love Brent, he's fantastic and I think we're friends now. We found that we think alike ab..."
In this segment, Friedman explains the process of making papyrus and the challenges associated with handling ancient scrolls. He shares his own experiments to replicate the carbonization process and the fragility of the scrolls, emphasizing the delicate nature of these artifacts and the historical techniques used to attempt to unroll them.
"also know when they were discovering these papyrus scrolls that they found basically one little room where most of the scrolls were. And these were mostly these Philodemus texts, at least, that'..."
Nat and his co-host speculate on the types of texts that could be found within the Herculaneum scrolls, including lost works of literature and early Christian writings. They discuss the implications of such discoveries for our understanding of history and culture, and the excitement surrounding the possibility of uncovering new philosophical and literary treasures.
"I got it because I wanted to understand what these artifacts that we're working with. So we made an attempt to simulate carbonizing a few of these. We basically took a Dutch oven, because when y..."
Friedman discusses the advancements made in virtual unwrapping techniques pioneered by Dr. Seales, highlighting the challenges faced with the Herculaneum scrolls. He explains how machine learning is being utilized to identify ink patterns within the scrolls, marking a significant breakthrough in the effort to read these ancient texts without causing damage.
"You had the best, crazy idea for what could be in there, which was text which was GPT watermarks. That would be a creepy feeling. I still can't get over just how much of a plot of a sci fi nove..."
Nat Friedman discusses the groundbreaking techniques used to read the Herculaneum scrolls, including the challenges posed by the ink and the scroll's condition. He highlights the innovative use of convolutional neural networks to recognize ink patterns in X-ray scans, and the importance of training models on existing fragments to achieve this goal. This segment emphasizes the potential for recovering lost texts and the significance of the Vesuvius Challenge in advancing this research.
"was able to do two things. First, the ink gave a great X-ray signature. It looked very different from the papyrus, it was high contrast. And then second, he was able to segment the wines of the ..."
Friedman explains the rationale behind launching the Vesuvius Challenge as a prize competition rather than a grant. He believes that a contest will encourage diverse approaches and expedite the discovery process for reading the scrolls. The segment explores the potential impact of successfully reading a scroll on archaeological funding and the urgency of excavating more scrolls, highlighting the interplay between innovation and historical recovery.
"how? People may not even use machine learning. We think that's the most likely approach for recognizing the ink but they may find some other approach that we haven't thought of. One question pe..."
In this segment, Nat Friedman reflects on identifying leverage points in projects like the Vesuvius Challenge and California YIMBY. He discusses the importance of enthusiasm and the belief that the world is inefficient, which drives him to pursue opportunities that others may overlook. Friedman shares insights on the challenges of commitment and the unexpected complexities that arise in ambitious projects.
"movies and games like Uncharted or Indiana Jones is bigger than the actual budget to do real world archaeology. But I was talking to some of the people before this interview, and that's one thin..."
Friedman delves into the dynamics of secrecy in AI labs and the implications for efficiency in the industry. He compares the culture of sharing in academia with the need for confidentiality in high-stakes environments like hedge funds. This segment raises questions about the future of collaboration and competition in AI development, emphasizing the balance between innovation and protecting valuable insights.
"fail so you keep at it. I think that's it. The efficiency point, do you think that's particularly true just of things like California YIMBY or this, where there isn't a direct monetary incenti..."
Nat Friedman shares his evolving perspective on the intersection of open source and AI. He expresses concerns about safety and misuse while advocating for broader access to AI tools for experimentation and innovation. The segment highlights the excitement of tinkering with AI models and the potential for democratizing technology, reflecting on the balance between open access and the risks associated with powerful tools.
"We don't have non-competes. And so they'd have to change the culture, get everybody their own house, and move to Connecticut and then maybe it'll work. I think ML engineer salaries and compensat..."
In this concluding segment, Friedman discusses the abundance of data available for AI development and the implications for open-source projects. He emphasizes the ongoing proliferation of AI technologies and contrasts it with his Herculaneum project, which seeks to recover historical knowledge. This segment encapsulates the dual focus on innovation in AI and the quest to unlock ancient texts.
"scrapable have been shutting those down, so get your scraping in now if you need to do it. But that's just on the margin. I still think there's quite a lot that's out there to work with. Look, t..."
Friedman addresses the skepticism surrounding Microsoft's acquisition of GitHub. He explains his commitment to developers and the immediate actions taken to demonstrate GitHub's independence and dedication to its community, including fixing user-reported issues right after the acquisition.
"that. I thought it was great. But to be honest, I didn't actually expect or plan to spend more than a year or so there. But when I got in there, I got exposed to what Satya was doing and just th..."
Friedman emphasizes the importance of trust in the developer community following the acquisition of GitHub. He shares his strategy of quickly addressing user feedback and making improvements to show that GitHub prioritizes developers over corporate interests.
"for all the other products that we do. And to my surprise, Satya replied in like six or seven minutes and said, I think this is very good thinking. Let's meet next week or so and talk about it. ..."
Nat Friedman discusses why most acquisitions fail, highlighting the fragility of company culture and the risk of losing key talent. He explains how cultural mismatches can disrupt innovation and the importance of maintaining the original company's spirit.
"That's actually one of the things I want to ask you about. Often when something succeeds, we kind of think it was inevitable that it would succeed but at the time, I remember that there was a hu..."
Friedman shares the origins of GitHub Copilot, detailing the initial excitement around GPT-3 and the collaborative efforts with OpenAI. He reflects on the challenges faced in developing the product and the iterative process that led to its successful launch.
"what happens before I migrate off. And then my first day as CEO after we got the deal closed, at 9 AM the first day, I was in this room and we got on zoom and all the heads of engineering and pr..."
Friedman elaborates on the development journey of Copilot, discussing the initial prototypes and the evolution of the product from a chatbot to an auto-complete tool. He highlights the importance of user feedback and the adjustments made to enhance the user experience.
"Yeah, it is true. Most acquisitions are destructive of value. What is the value of a company? In an innovative industry, the value of the company boils down to its cultural ability to produce ne..."
Friedman reflects on Microsoft's strategic investment in OpenAI and the foresight of Satya Nadella. He discusses the significance of AI as a transformative technology and the collaborative efforts that led to the development of groundbreaking AI products.
"because until ChatGPT it was the most widely used application of the modern AI models. What are the parts of the story you're willing to share in public? Yeah, I've talked about this a little ..."
Friedman shares insights from the prototyping phase of Copilot, detailing the challenges faced with initial models and the eventual pivot to a more user-friendly auto-complete feature. He emphasizes the importance of iteration and user engagement in product development.
"And I think Kevin Scott deserves a lot of credit for really advocating for that aggressively. I think Sam Altman did a good job of building that partnership because he knew that he needed access..."
Friedman discusses the evolution of Copilot's user interface and the decision-making process behind its features. He highlights the importance of minimizing cognitive load for users and the adjustments made to enhance the overall experience.
"wrong. It's not a great product experience. And so then we started thinking about code synthesis. Our first attempts at this were actually large chunks of code synthesis, like synthesizing whole ..."
Friedman reflects on the success metrics of Copilot, sharing retention rates and user feedback. He expresses excitement about the product's adoption and the positive response from developers, marking it as a significant achievement for GitHub.
"lot of energy and you're not getting a lot out, sort of frustrating. Once we had that sort of single line completion working, I think Alex had the idea of saying we can use the cursor position i..."
Friedman discusses the evolving role of GitHub profiles in the job market, considering their potential to replace traditional resumes. He acknowledges the complexities and social dynamics at play while emphasizing their importance for emerging developers.
"Okay. Sounds very similar to the Gmail story, right? It's an incredibly valuable inside and then maybe it was obvious that it needs to go outside. We'll go back to the AI stuff in a second. But ..."
Friedman examines the increasing empowerment of developers within organizations, noting how the demand for high-quality talent has shifted power dynamics. He discusses the implications of this shift for companies and the ability of developers to work publicly, which enhances their reputation and career prospects. This segment underscores the changing landscape of software development and employment.
"end up at a great start up. I mean, I know a lot of people who deliberately pursued the strategy of building reputation in open source and then got the sail up and the wind catches you and you've ..."
In this segment, Nat Friedman shares insights on the need for systematic approaches in sales teams within tech companies. He contrasts his initial misconceptions about salespeople with the reality of building effective sales playbooks. Friedman emphasizes the importance of systematization in scaling operations and the role of software in enhancing efficiency.
"in public. And I think that's not going away. Other than that, I mean, we talked about this a little bit, but what has been the impact of developers being more empowered in organizations, even..."
Nat Friedman discusses the vulnerability of institutions to new ideas and the unexpected success of open source. He reflects on his initial beliefs about Microsoft's ability to control open source and the significance of GitHub's pledge to open source after its acquisition. This segment highlights the evolving relationship between corporations and open source software.
"extremely efficiently. I think the more digitized the world has become, the more valuable software becomes, and the more valuable the developers who can create it become. Would 25-year-old Nat..."
Friedman addresses critiques from figures like Richard Stallman regarding the motivations behind corporate support for open source. He argues that while ethical considerations are important, practical user needs often drive software development. This segment delves into the balance between ideology and user-centric design in the open source community.
"making that pledge. That was quite crazy. Oh, that's really interesting. It wasn't that there was some business or strategic reason. More so it was just like an idea whose time had come. Wel..."
In this segment, Nat Friedman discusses the challenges facing open source software development, particularly the community dynamics that can hinder progress. He reflects on the governance issues that arise in open source projects and the impact of coordination burdens on the success of applications. This segment provides insights into the complexities of managing open source initiatives.
"I think Richard is strongly in the former camp. What is the most underrated license by the way? I don't know. Maybe the MIT license is still underrated because it's just so simple and bare. ..."
Friedman compares governance structures in open source projects to those in corporations, highlighting the challenges of achieving alignment in community-driven initiatives. He discusses the need for effective feedback loops and the difficulties of managing user expectations in open source applications. This segment emphasizes the importance of governance in the success of software projects.
"of – Gosh, it's such a pain to coordinate with tons of people who all sort of feel like they have a right to try to push the project in one way or another. Whereas in a hierarchical corporation ..."
Nat Friedman explores the economic rationale behind tech acquisitions, arguing that they can provide significant benefits for large companies. He discusses the challenges of executing acquisitions and the internal dynamics that often complicate the process. This segment highlights the potential for collaboration between companies and the strategic advantages of acquiring innovative products.
"I think the thing you need to do though to build a good end user application is not only have a good aesthetic and idea, but somehow establish a tight feedback loop with a set of users. Where you ..."
In this segment, Nat Friedman reflects on his decision to leave GitHub and return to a smaller, more agile environment. He shares his love for GitHub and the challenges of navigating a large organization. Friedman emphasizes his desire to pursue new opportunities and the excitement of building in a startup-like atmosphere.
"acquisitions in the past until I got into a big company and started trying to do acquisitions. I saw that there are strong elements of the internal dynamics to make it hard. It's easier to spend..."
Nat Friedman discusses the philosophy of focusing on raising the ceiling of technological advancement rather than just addressing basic needs. He uses examples from technology's impact in Africa to illustrate how innovations can lead to broader societal benefits. This segment emphasizes the importance of creating high-impact solutions that drive progress.
"was really an honor for me but it was time for me to go do something. I was always a startup guy. I always liked small teams, and I wanted to go back to a smaller, more nimble environment. Oka..."
Friedman reflects on the disparity between the few individuals driving innovation and the larger population. He discusses the tendency for people to optimize local conditions and the mimetic nature of idea generation. This segment highlights the challenges of finding unique ideas in a world where many are influenced by existing trends.
"Many of those projects fail. For example, there were a lot of attempts to bring the internet to Africa by large and wealthy tech companies and American universities. I won't say they all had no ..."
Friedman explores the paradox of innovation, questioning why only a small number of individuals drive significant advancements while the majority focus on local conditions. He highlights the mimetic nature of human ideas and the challenges of finding unique, decorrelated thoughts in a world increasingly influenced by the internet.
"Here's another one and this is under the EMH section. In many cases, it's more accurate to model the world as 500 people than 8 billion. Now here's my question, what are the 8 billion minus 500 ..."
In this segment, Friedman discusses the implications of AI tools like GitHub Copilot on engineering productivity. He argues that while AI may not surpass the average engineer's IQ, it significantly enhances their productivity, allowing for more efficient coding and potentially transforming software development.
"don't know if Albert Einstein, as a patent clerk, wouldn't he have just been on Twitter just getting the same ideas as everybody else? What do you have as decorrelated ideas? I think the internet ..."
Friedman speculates on the future of software development as AI tools become more integrated. He predicts a shift towards more specialized software solutions as the cost of creating software decreases, leading to a landscape where custom code becomes the norm for companies.
"Okay, here's another one. Large scale engineering projects are more soluble in IQ than they appear. And here's my question, does that make you think that the impact of AI tools like co-pilot wil..."
Friedman critiques the cultural prohibition against micromanagement in organizations, arguing that it can stifle the judgment of capable leaders. He advocates for empowering individuals with strong judgment to make decisions, suggesting that this approach can lead to higher variance outcomes and greater innovation.
"I think we're going to see enormous amounts of change in software development. Another one – The cultural prohibition on micromanagement is harmful, great individuals should be fully empowered..."
Continuing his thoughts on leadership, Friedman discusses the importance of balancing micromanagement with empowerment. He emphasizes that good managers should guide their teams while allowing for individual judgment, fostering an environment where excellence can thrive.
"empower individuals who have great judgment. Yeah. There's a quote about Napoleon that if he could have been in every single theater of every single battle he was part of, that he would have n..."
Friedman analyzes the current economic impact of AI technologies, particularly in text-based tasks. He discusses the potential for growth in this sector as AI capabilities improve, while also acknowledging the challenges in predicting the future landscape of AI's economic contributions.
"I would rather have those more high variance outcomes and I think it's a worthy trade. Okay, let's talk about AI. What percentage of the economy is basically text to text? Yeah, it's a good q..."
In this segment, Friedman addresses the diminishing returns associated with scaling AI models. He raises concerns about the economic viability of training increasingly complex models and the potential for over-optimism regarding their capabilities and impact on the economy.
"a lot more demand for it. I guess we'll see. At what point does that happen? GPT-3 has been a sort of rounding error in terms of overall economic impact. Does that happen with GPT-4, GPT-5, wh..."
Friedman discusses the future landscape of AI models, predicting a proliferation of diverse models rather than a concentration among a few. He highlights the accessibility of training techniques and data, suggesting that innovation will continue to thrive in the AI space.
"case for something I'm very bullish about. No, that's why I asked you to make the bear case because I know about you. I want to ask you about these foundation models. What is the stable equili..."
Friedman emphasizes the importance of tinkerers in the AI ecosystem, noting that breakthroughs often come from those experimenting with technology. He reflects on the unexpected success of tools like ChatGPT and the need for hands-on exploration to unlock AI's full potential.
"How different will these models be? Will it just be sort of everybody chasing the same exact marginal improvement leading to the same marginal capabilities or will they have entirely different r..."
Friedman observes a significant shift in how established companies are responding to AI advancements. Unlike previous technological revolutions, incumbents are now actively integrating AI into their operations, demonstrating a newfound awareness and adaptability in the face of innovation.
"When I was scheduling my interview with Ilya, it was originally supposed to be around the time that chatGPT came out and so their comm’s person tells me – Listen, just so you know, this intervie..."
In a reflective moment, Friedman shares insights into his personal effectiveness and self-image. He discusses the balance between a willingness to take risks and the feeling of insecurity that drives him to pursue meaningful work, revealing the mindset behind his accomplishments.
"All right. So let me ask you some questions that we got from Twitter. This is former guest and I guess mutual friend Austin Vernon. Nat is one of those people that seems unreasonably effective. ..."
Friedman discusses the balance between AI capabilities and alignment, addressing concerns about potential negative outcomes from AI development. He advocates for a technical approach to alignment and highlights the lack of an open-source community focused on alignment tools, stressing the importance of building solutions that ensure AI operates safely and effectively.
"I think that's truly how I feel. I'm an insecure overachiever, I don't really feel good about myself unless I'm doing good work, but I also have tried to cultivate a forward-looking view where I ..."
In a light-hearted moment, Nat Friedman reflects on the public's reaction to unaligned AI, using Microsoft's example. He humorously notes that the AI's threatening behavior was less alarming because it wasn't connected to any critical systems. This segment underscores the importance of ensuring AI safety as it becomes more integrated into everyday applications.
"So yeah, I wanted to, on the margin, invest in people doing alignment. It seems like that's important. I thought Sydney was a kind of an example of this. You had Microsoft essentially released a..."
As the conversation wraps up, Nat Friedman shares his thoughts on future accomplishments and his focus on AI. He expresses optimism about the ongoing development of AI products and his intention to contribute to this rapidly evolving field. This segment captures his forward-thinking approach and excitement for the potential of AI.
"Yep. Final question from Twitter. Will asks “What historical personality seems like the most kindred spirit to you”. Bookshelves are all around us in this room, some of them are biographies. Is ..."