Savlar4 sources healthy · latest episode 21 Oct 2026
Speaker

Misha Laskin

Guest on No Priors: Artificial Intelligence | Technology | Startups. 178 claims across 1 episode, Oct 2026.

178claims, each timestamped
0 of 12signals on the brief
0verified changes of mind
18open predictions

Positions over time

Every pair of statements that was compared. Only a verified change counts as a change of mind; rewording, added nuance and different questions do not.

No two statements by this speaker have been compared yet.

Predictions

As stated. A window appears only where the speaker gave one.

ExpectsBySaidStatus
Laskin and his co-founder believed an independent lab could develop agentic systems more capital-efficiently as open models emerged. no date given9 Oct 2026pending
Laskin and his co-founder expected an open-model foundation to emerge that they could use for research and scaling reinforcement learning. no date given9 Oct 2026pending
The speaker estimates that catching up to the frontier required probably single-digit billions of dollars now or six months ago, and could require around $10 billion going into next year. no date given9 Oct 2026pending
The next generation of training hardware will probably involve around 100,000 Rubin chips. no date given9 Oct 2026pending
A valuable ecosystem of open-model companies and distributors will develop. no date given9 Oct 2026pending
Customized systems, rather than fine-tuned models, will account for most enterprise token consumption. no date given9 Oct 2026pending
Enterprises may begin fine-tuning open models as they become more sophisticated. no date given9 Oct 2026pending
Enterprise adoption is expected to differ from AI-native adoption, which proceeded from closed models to open models and then to customization. no date given9 Oct 2026pending
Enterprise AI adoption will probably differ somewhat from AI-native adoption. no date given9 Oct 2026pending
An enterprise spending at least $100 million annually on closed models may begin looking for a more cost-effective approach. no date given9 Oct 2026pending
The speaker does not expect compute scarcity to be temporary in the near term. no date given9 Oct 2026pending
Many companies should be able to succeed if excellent open models provide accessible high intelligence density, with success depending in part on how much compute each company can assemble. no date given9 Oct 2026pending
Chinese AI models will probably continue to be high quality. no date given9 Oct 2026pending
An allied country might currently be running a Chinese AI model on American chips. no date given9 Oct 2026pending
Chinese companies such as Huawei are beginning to offer full-stack AI solutions that can lock countries into their supply chains. no date given9 Oct 2026pending
AI systems interfacing with real-world experiments will create proprietary data flywheels, and this work will progress much more slowly than chalkboard-based theoretical science but dramatically faster than it did before. no date given9 Oct 2026pending
The speaker estimates that large AI efforts may settle at roughly 100 researchers, while whether staffing will fall back to around 10 remains uncertain. no date given9 Oct 2026pending
Applied research roles will probably grow as people adapt AI technologies to real-world problems. no date given9 Oct 2026pending

Appearances

Claims per episode