Meta Superintelligence Labs VP of Product & Applied Research
Nat Friedman
Profile
Nat Friedman is the person who shipped the product that proved AI coding assistants were real. As CEO of GitHub from October 2018 through November 2021 — the post-Microsoft acquisition era — he oversaw the launch of GitHub Copilot in June 2021, built in partnership with OpenAI. At the time it felt like a novelty demo. Four years later, every IDE, coding agent, and “AI-native” developer tool on the market is chasing the template he set.
Before GitHub, Friedman spent two decades in open-source infrastructure. He co-founded Ximian (GNOME desktop tooling) with Miguel de Icaza in 1999, then Xamarin for cross-platform mobile dev, which Microsoft acquired in 2016. That arc — OSS → developer platforms → AI tooling — is why he’s worth paying attention to. He’s been near every major shift in how developers actually work for 25 years, and he recognizes good leverage when he sees it.
After leaving GitHub, Friedman and Daniel Gross built one of the most active AI investment operations in the industry. NFDG took early stakes in Safe Superintelligence, Perplexity, Character.ai, and Midjourney (where he’s also an advisor). Their AI Grant accelerator writes $250K checks plus compute credits to seed-stage AI startups — the money matters less than the signal. He also co-founded the Vesuvius Challenge with Brent Seales, which in 2023–2024 successfully used machine learning to read the carbonized Herculaneum scrolls for the first time in 2,000 years — a rare case of crowdsourced AI being used on a problem that actually mattered outside tech.
In June 2025, Meta effectively acqui-hired Friedman and Gross through a deal that bought out a large portion of NFDG’s LPs. Friedman joined Meta Superintelligence Labs as VP of Product and Applied Research, reporting to Alexandr Wang. He now shapes how Llama research turns into actual consumer products. In parallel, he’s quietly started a new developer-tools company, Entire, which raised $60M seed in early 2026. If you’re building AI developer tools today, Friedman is simultaneously the investor you want, the operator you should study, and — increasingly — a competitor.
Key Articles & Papers
Introducing GitHub Copilot: your AI pair programmer Vesuvius Challenge: First Word Discovered in Unopened Herculaneum Scroll Stratechery: An Interview with Daniel Gross and Nat Friedman about the Democratization of AI Stratechery: Reasoning About AI Stratechery: Models, Margins, and Moats Nat Friedman's personal site AI GrantControversies
Copilot and open-source licensing. At Copilot’s launch, Friedman publicly argued that “training ML systems on public data is fair use.” That position became the center of Doe v. GitHub (2022), a class action alleging Copilot reproduced GPL-licensed code without attribution. Most claims were eventually dismissed, but the case framed the broader debate about training on scraped code that’s still playing out across the industry. Background.
Meta’s NFDG deal. Meta’s mid-2025 buyout of NFDG LP stakes to bring Friedman and Gross in-house was widely read as a regulatory-friendly acqui-hire — the same pattern as Microsoft/Inflection and Amazon/Adept. Critics argued it absorbs AI talent and portfolio influence without triggering antitrust review. Coverage.
YouTube
Spotify Podcasts
