Unsupervised Learning with Jacob Effron
Unsupervised Learning with Jacob Effron

Ep 92: xAI Co-Founder Unpacks the Future of Model Development

July 31, 2026

AI Summary

5 min read

In December 2024, Igor Babushkin—former co-founder of xAI and a key figure at DeepMind and OpenAI—spent part of the month rewriting every coding project he had ever wanted to do, using coding agents that had suddenly become powerful enough to make it feel like "anything is possible." That moment crystallized something for him: the same transformation that hit software engineering is about to hit everything else. "All of computing really is up for grabs when it comes to agents," he says. The episode is a conversation about what comes next—how to build agents that work in non-verifiable domains, why the closed-source model providers are getting squeezed, and what it means to give individuals and companies real control over their AI.

The Next Frontier: Beyond Verifiable Domains

Babushkin argues that the current wave of progress has been driven by domains with clear, verifiable rewards—coding and math, where you can check whether a program compiles or a proof holds. This has been the "central path" and the low-hanging fruit. But the next big leap requires moving into domains where the reward signal is fuzzy: scientific discovery, material science, or even personal productivity.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (00:01) **Introduction and Guest Background** - Jacob introduces Igor Babushkin, co-founder of xAI, and his new company River AI.
  • 2 (01:53) **The Inspiration Behind Igor's Fiction** - Igor discusses two short stories he wrote, exploring the implications of powerful AI agents.
  • 3 (03:59) **The Next Frontier Beyond Coding** - Igor identifies the key challenges and opportunities for moving AI agents into non-verifiable domains.
  • 4 (09:23) **The Core Thesis of River AI: Democratizing Control** - Igor explains the founding philosophy behind his new company, River AI.
  • 5 (12:13) **River AI's Three Major Bets** - Igor outlines the three distinct projects River AI is pursuing.
  • 6 (16:36) **The Case for Local, Private AI** - Igor elaborates on the benefits and challenges of running powerful AI on local hardware.
  • 7 (18:27) **Training for Individual Happiness** - Igor discusses the research approach to creating truly personalized AI agents.

+ Full timestamped outline available in the app

Show Notes

Igor Babuschkin, co-founder of River AI and formerly a co-founder of xAI, joins to unpack a career that spans nearly every major AI lab: he led the StarCraft and AlphaCode work at DeepMind, joined OpenAI's reasoning team years before o1 shipped, and co-founded xAI, where he helped stand up the Colossus data center in roughly 120 days and reflects candidly on what it's actually like working with Elon Musk day to day, plus what the Cursor acquisition actually unlocked for Grok's coding models. He also discusses why he left xAI to start River AI, the three bets behind it, and why he's betting on local hardware, not just software, for personal AI. On the enterprise side, he tackles whether companies will actually train their own models or if it's just a cost play, and makes the case that proprietary labs like OpenAI and Anthropic are facing a real business squeeze. He's skeptical that stacking specialized RL domains generalizes the way pre-training scale did, and is candid about the uncomfortable reality that today's frontier open-weight models are almost entirely Chinese. He closes on what's actually needed to push model progress beyond coding into non-verifiable domains, and the broader implications of where AI is headed next.

 

(0:00) Intro
(1:17) Writing Fiction on Where AI Is Headed
(4:46) Cracking Agents Beyond Coding
(10:29) Why Igor Left to Start River
(12:22) River's Three Big Bets
(18:06) Weights vs. Memory: The Personalization Debate
(22:04) Should Enterprises Train Their Own Models?
(25:10) Are Proprietary Labs Losing Their Edge?
(32:16) The China Open-Source Problem
(44:19) The Elon Call That Started xAI
(50:18) Thoughts on Cursor Acquisition
(52:16) What's Actually Bottlenecking AI
(56:55) Humans, Machines, and Staying Relevant
(1:01:29) Igor's Odds This All Goes Well

 

With your host:  

@jacobeffron  

- Managing Director at Redpoint

Unsupervised Learning with Jacob Effron

More from this podcast

Unsupervised Learning with Jacob Effron →