{"version":"1.0","type":"rich","provider_name":"Yank","provider_url":"https://yank.info","title":"“Pre-training is full of researchers who chew glass and stare out into the abyss.” — Techno Optimist","author_name":"Techno Optimist","html":"<blockquote class=\"yank-embed\" data-yank=\"zuck-and-wang-on-the-art-1let7u/1\" cite=\"https://yank.info/gr4uv5/1\">\n  <p>&ldquo;Pre-training is full of researchers who chew glass and stare out into the abyss.&rdquo;</p>\n  <footer>&mdash; Techno Optimist, <cite>Techno Optimist on Instagram: &quot;Zuck and Wang on how to pre-train an AI model:\n\nWang: “In the pre-training stage, everyone who’s training frontier models basically is trying to get as much data as poss</cite> &middot; <a href=\"https://yank.info/gr4uv5/1\">via yank.info</a></footer>\n</blockquote>\n<script async src=\"https://yank.info/embed.js\" charset=\"utf-8\"></script>","width":640,"height":240,"cache_age":86400}