Coordinal: A Postmortem. by Ronak Mehta

·Nuno Sempere··

cart;horse: Most star­tups fail, this one did too. Tried too much with too lit­tle, with a lit­tle too much ego.Co­or­di­nal Re­search’s goal was to build an au­to­mated safety re­search plat­form. A re­searcher writes “Repli­cate X re­sult from pa­per Y with tweak Z” and the sys­tem pro­vi­sions a sand­boxed com­pute en­vi­ron­ment, gath­ers con­text, writes code, runs ex­per­i­ments, and re­turns a re­search re­port plus over­sight trail.The bet was that AI-ac­cel­er­ated re­search is com­ing ...

Read full article →

Related Articles

Landmark new METR report: Can AIs already start ‘rogue deployments’ inside AI companies? by 80000_Hours
80000_Hours · Nuno Sempere · 22h ago
Will the next full gemini model be frontier at coding?
Ian Shea · Manifold Markets · 1d ago
Will there be more than 100 cases of ebola in the US in 2026?
Joseph Caissie · Manifold Markets · 2d ago
Robots reliably do my laundry by?
Mochi · Manifold Markets · 2d ago
Will the next full gemini model be as good as opus 4.7 or gpt 5.5 at coding?
Ian Shea · Manifold Markets · 2d ago