{"id":22681,"date":"2026-07-23T05:09:14","date_gmt":"2026-07-23T05:09:14","guid":{"rendered":"https:\/\/scannn.com\/inside-the-model-factory-eiso-kant-poolside-ai\/"},"modified":"2026-07-23T05:09:14","modified_gmt":"2026-07-23T05:09:14","slug":"inside-the-model-factory-eiso-kant-poolside-ai","status":"publish","type":"post","link":"https:\/\/scannn.com\/lv\/inside-the-model-factory-eiso-kant-poolside-ai\/","title":{"rendered":"Inside the Model Factory \u2014 Eiso Kant, Poolside AI"},"content":{"rendered":"\n
\n

In recent months, the open vs closed, and <\/span>US vs China<\/a> discussions on model ownership and sovereign\/local AI have heated up to a fever pitch. So it is very very good news that <\/span>Poolside AI<\/a><\/strong> are finally emerging with new models, like <\/span>Laguna S 2.1<\/a><\/strong>, that are <\/span>beating Thinking Machines\u2019 recent release nearly 10 times their size<\/a>.<\/span><\/p>\n

\n
<\/p>\n
<\/picture><\/div>\n

<\/a><\/figure>\n<\/div>\n

Poolside\u2019s recent tech report<\/a> got a lot of praise due to their level of detail, and Vibhu first covered Laguna\u2019s recent technical report on our paper club:<\/span><\/p>\n

<\/p>\n

\n
\n
\n
\"X<\/picture><\/div>\n<\/div>\n

Eiso Kant<\/span>@eisokant<\/span><\/p>\n

<\/g><\/svg><\/div>\n

Loving the @latentspacepod<\/span> breakdown of our Laguna M.1\/XS.2 Technical Report! The Latent Space paper club just did a deep dive, and their takeaways perfectly capture what we set out to build with our Model Factory. A few quotes from the video (1\/6)
\nyoutu.be\/QLfZamyMls0<\/a><\/div>\n
\n

8:34 PM \u00b7 May 28, 2026<\/span> \u00b7 <\/span>11.3K Views<\/span><\/p>\n

1 Reply<\/span> \u00b7 <\/span>8 Reposts<\/span> \u00b7 <\/span>47 Likes<\/span><\/p>\n<\/div>\n<\/div>\n

<\/a><\/p>\n

From spending <\/span>$12 million building language models<\/strong> for code before the world cared to <\/span>creating a Model Factory<\/strong> that can take a model from pre-training to release in <\/span>eight weeks<\/strong>, Eiso Kant has spent more than a decade <\/span>betting that code is the path to AGI<\/strong>. In this episode, the Poolside co-founder joins swyx and Vibhu to explain why ChatGPT felt like vindication, why Poolside embraced open weights and open research, and why he <\/span>would rather live in a world with 100 foundation model companies than five even if Poolside were one of the five.<\/strong><\/p>\n

\n