Zuckerberg says Llama 4 failed because it was staffed like Instagram, not a frontier lab
rohanpaul_ai · x · 2026-10-06
In a interview on The Next Big Thing YouTube channel, Mark Zuckerberg explained why the Llama run broke after Llama 3: the model was near the frontier, but Llama 4 fell off the path. His diagnosis is organizational — Meta staffed the effort like Instagram feed and ads, with hundreds or thousands of people working in parallel, when frontier model research requires a small, tight science team.
He also reflected on entrepreneurship: things inevitably go wrong, and what defines the trajectory is how you figure out how to move forward when they do.
More from Companies & People
- OpenAI shares revenue with popular plugins, model picker may go away, says ChatGPT head — gaganghotra_ · 2026-10-06
- Sakana AI hiring Public Sector Associate to manage government-funded projects — SakanaAILabs · 2026-10-06
- Anton Osika: 'It's hard to build in Europe, do it anyway' — syhw · 2026-10-06
- Julien Simon: Calling Aleph Alpha's Kolibri 'Europe's DeepSeek Moment' Is a Misreading — julsimon · 2026-10-06
- Karpathy's arrival makes Anthropic's coders better, says Kuprel — Kuprel · 2026-10-06
- B2E AI Needs Mission Design, Not Personas: UX Magazine on Enterprise AI — uxmag · 2026-10-06