Audio · 2026-07-23 · 6 moments

Inside the Model Factory — Eiso Kant, Poolside AI

Poolside's co-CEO on how his small team of top researchers built a model factory capable of training Laguna S - a 118B MOE beating Thinky's ~1T open weights model... and this is just the beginning. ✦ AI generated

Apple Podcasts ↗

GUEST Eiso Kant HOST Alessio Fanelli HOST Shawn Wang (swyx)

timeline · colored by role

Context◆

The world would be better off with 100 foundation model companies than an oligopoly of five, even if Poolside were one of the five.

Eiso explains Poolside's pivot to open source: he believes intelligence will become the world's most demanded commodity, and that an oligopoly of a few companies controlling all intelligence would be dystopian. He argues that regulation risks accidentally locking in an oligopoly, as happened with cigarette advertising bans.

transcript

Eiso Kant: I rather live in a world that has 100 foundation model companies than a world that has five, even if I was one of the five. And the smallest and most meaningful contribution we can make for 100 to exist is to open up our research and open up our weights right now.

Claim◆

Model building is ultimately 90% engineering — it's an industrialized process of data pipelines, distributed training, and reproducibility, not pure research magic.

Eiso argues that model building is an engineering discipline, not theoretical magic. The Model Factory treats training as an end-to-end industrial process with streaming data, immutable data layers, and versioned code.

transcript

Eiso Kant: I would say that our view from very early on in the company was that model building is ultimately 90% engineering. And I think we all know it in the industry because if you look at where's every researcher spending their time, they're spending their time writing code, right? Looking at data and writing code.

Definition

95% of model building reduces to improving data or improving compute efficiency — it's not magic, and many more people could be capable researchers.

Eiso demystifies foundation model work: most of the job is looking at data, writing good code, and running experiments. He cites an example of an engineer who became a legitimate RL researcher in six months because the Model Factory lowered the experimentation barrier.

transcript

Eiso Kant: I think you can sum down 95% of model building to just doing two things. You're improving data or you're improving compute efficiency... A lot of the work that all of us are doing on a day is not sitting down trying to solve a math theorem. A lot of the work that we're doing is just really doing the basics right, writing good code, looking at data, improving it, running experiments.

extends · 1

Mechanism◆

Smaller models can be far more capable than expected if trained for persistence, verification, and backtracking behaviors rather than raw intelligence alone.

Eiso explains that Laguna S (118B total, 8B active) outperforms models many times its size because post-training instilled behaviors like persistence and backtracking, not just raw intelligence. He suggests this means the peak of model usefulness for knowledge work may be much closer than previously thought.

transcript

Eiso Kant: The gains in Laguna S come not from more intelligence, but more from different behavior, more verification, less taking things for granted, not declaring victory early, and being way more persistent... This model for me is the first sign that maybe that peak is at a trillion, five trillion, ten trillion. Maybe we can just squeeze way more out of these models.

provides context · 1supports · 1

Prediction

Reinforcement learning will move earlier and earlier into pre-training, and next-token prediction alone is not extracting enough from the web.

Eiso believes RL will shift earlier into the training process, and that current next-token prediction is leaving most of the web's learning potential untapped. He views distillation and synthetic environments as 'drugs' the industry is addicted to, and thinks the real gains will come from teaching models to think earlier in training.

transcript

Eiso Kant: I have a not commonly held opinion that reinforcement learning will move earlier and earlier into training. We've been spending a couple of years really doing research on how can we turn the web into not just next token prediction, but into a way to teach the model to think earlier in its training. And I think there's a huge amount of gold to be found there.

provides context · 1

Claim

MCP and traditional tool calls are 'stupid' — the future is models writing code freely in a minimal container, not choosing from 50 predefined tools in a system prompt.

Eiso argues that stuffing 50 tools into a system prompt is the wrong direction. Instead, models should be given a minimal harness with a container, codebase, and API keys, and allowed to write scripts with if-statements and loops to solve complex tasks. He predicts this will be the norm within 12 months.

transcript

Eiso Kant: I think MCP and tools are stupid... we are moving from tool calls to effectively models writing code, little scripts... Give the model a minimal harness, as minimal as possible, give it a container in which it has its own code base... and just let it run free at the task. And I think that is the way we're going.

rebuts · 2

Highlight slides

The case for 100 foundation model companies over 5✦ from: The world would be better off with 100 foundation model companies than an oligopoly of five, even if Poolside were one of the five.Poolside's open-source bet✦ from: The world would be better off with 100 foundation model companies than an oligopoly of five, even if Poolside were one of the five.Model Building Is 90% Engineering✦ from: Model building is ultimately 90% engineering — it's an industrialized process of data pipelines, distributed training, and reproducibility, not pure research magic.The Model Factory Mindset✦ from: Model building is ultimately 90% engineering — it's an industrialized process of data pipelines, distributed training, and reproducibility, not pure research magic.Laguna S: Small Model, Big Results✦ from: Smaller models can be far more capable than expected if trained for persistence, verification, and backtracking behaviors rather than raw intelligence alone.The Knowledge-Work Peak May Be Closer Than Expected✦ from: Smaller models can be far more capable than expected if trained for persistence, verification, and backtracking behaviors rather than raw intelligence alone.

Related episodes

Video · All-In Podcast

The Fight Over Open Source AI, Anthropic's $1.5B Payout, NYC Socialists: Evictions = Violence?

→

Video · The TWIML AI Podcast

Teaching Computers to Smell | Alex Wiltschko

→

Video · The Cognitive Revolution

LTX CEO on Video Gen + Can we beat AI superforecasters?

→