Eiso Kant
- 1 episode on Chain of Thought
- 1 topic covered
- 1 full transcript
- May 2025 first on the show
Eiso Kant is the co-founder and CTO of Poolside, an AGI company focused on code. On Chain of Thought he and co-founder Jason Warner laid out Poolside’s bet on reinforcement learning and deep specialization for coding as a route to AGI.
What this guest adds
Eiso Kant argues that even as AI adds to the supply of software development, there is still much more software to build. Read transcript at 43:24
Episodes
What the conversations cover
These notes summarize the episodes featuring Eiso Kant, including other speakers in panel conversations. Follow a source link to read the discussion in context.
EP 23 First Code, Then AGI: Software’s Event Horizon with Poolside Founders Jason Warner & Eiso Kant
- Poolside is an AGI company, and code is the path it chose. Jason Warner: “Poolside is an AGI company. We’re going after AGI because we wanna build intelligence on compute.” Code is the chosen path — a “proxy task for intelligence,” Eiso Kant argues, because it demands both broad world knowledge and long-horizon reasoning. Episode takeaways
- The 2023 contrarian bet was reinforcement learning, not bigger pretraining. The consensus then: just “scale up next token prediction, scale up model size, provide more data, and we are, quote unquote AGI.” The Poolside founders bet the key scaling axis was RL instead — and “a lot of people thought we were kind of crazy.” In ’25, Jason says, it “looks more than prescient.” Episode takeaways
- Specialization is a budget decision. About a third of the way into training, the Poolside founders say, “we start biasing them massively towards software development.” Early on the models look general-purpose — “great at planning your trip to London or writing a poem.” Flip “a couple of dials” and they’d become much more like general-purpose models; the choice is where to spend finite compute. Episode takeaways
In their words
Once we are through the things that drive the cost of goods and services down to their raw material cost, we will start increasingly more and more spending on the frontier science where we don't know how much intelligence it takes. And we will also start building, taking this first primitive of intelligence as human intelligence and really applying it to very specific areas.
Eiso Kant First Code, Then AGI: Software’s Event Horizon with Poolside Founders Jason Warner & Eiso Kant · 38:15 Transcript excerpts; ellipses mark omitted words. Some automatically selected quotes have filler words removed.