Andrew White 🐦‍⬛ ✓

@andrewwhite01 · San Francisco, CA · joined 14 Apr 2009

Automating science. Cofounder @EdisonSci. Cofounder @FutureHouseSF. Former prof of chem eng.

33 567Followers

1 858Following

7 030Posts total

143.2KViews on collected posts

Ultimi post

Andrew White 🐦‍⬛ ✓@andrewwhite012 Sep

Look at those science evals! Great LABBench2 and biomystery bench scores! 4.3K views · 37 likes · 4 reposts · 2 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0126 Aug

Bixbench3: Can models reproduce the entire analysis from papers? We had agents repro papers end-to-end, with over 1B tokens spent on some. Frontier agents still score below 50%, and due to real problems like environment misconfig, quitting, or faking data 1/6 https://t.co/AYa6pDX 23.1K views · 296 likes · 43 reposts · 15 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0125 Aug

Someone is putting up hand-drawn flyers about crows in my neighborhood. I really like it https://t.co/dr2ndUIcnO 4.2K views · 62 likes · 3 reposts · 6 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0124 Aug

@andrewwhite01 @nattyover Yes and no. My experience is that the human spending more time than this aiding the output quality, pertinence, clarity, generalizability etc. Is almost always necessary. But the basic point is sound. I’d recast it as st approaching more and more that AI 203 views · 1 likes · 0 reposts · 0 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0124 Aug

@andrewwhite01 There's also like people trying to make a fully autonomous scientist but I like the kinda flexibility I get with Claude Code and Codex where I can "automate my way of doing science" :) 843 views · 1 likes · 0 reposts · 0 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0123 Aug

@andrewwhite01 Things that look deep can be shallow; a lot of ‘science’ that was shallow can be done by a LLM today, & whilst it was not deep it can help train students to think. The mistake here is be excited about shallow results that an AI can clearly do & take the training aw 2.1K views · 12 likes · 0 reposts · 1 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0123 Aug

Gavin Crooks is arguably the #1 researcher in the field of non-equilibrium stat mech and he just posted a major result from his phone on X while traveling as the result of some back and forth with Claude Fable. We are truly in a new world 93.1K views · 393 likes · 43 reposts · 9 replies Open on X →

Andrew White 🐦‍⬛ ✓@andrewwhite0131 Jul

We've released a new open-weights vLLM for reading chemical structure patents. This is a hard problem. These "Markush" structures are like graph regex: patterns that match many molecules. We used autoresearch and I expect many new models like this to emerge soon in science 1/4 ht 15.4K views · 202 likes · 40 reposts · 9 replies Open on X →

Rispetto ad account della stessa dimensione

7 post degli ultimi 90 giorni, accanto alla fascia di 10K–100K follower. arriva a molti, ma pochi di loro reagiscono.

Visualizzazioni mediane4 173questo account924mediana per 10K–100K

Copertura, %12.43%questo account3.62%mediana per 10K–100K

Interazione, %0.63%questo account1.52%mediana per 10K–100K

Metrica	Questo account	Mediana per 10K–100K	Rapporto
Visualizzazioni mediane per post	4 173	924	4.52×
Copertura (visualizzazioni ÷ follower)	12.43%	3.62%	3.43×
Tasso di interazione	0.63%	1.52%	0.42×

Altri account di questa fascia → Confronta con un altro account → Come sono costruiti questi parametri →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.