Civilization VC retweetledi
Tahoe’s vision is becoming real! It is exciting to see @iamjohnnyyu, @nalidoust, and the @tahoe_ai team bring years of focused building to this point.
I have watched firsthand how many pieces had to come together to make this possible. This team took the platform we built in the lab and transformed it into something that operates at a fundamentally different scale. Tahoe-100M was an early manifestation of that effort, and it is the part most people know. What is less visible is the enormous amount of data the team has generated across increasingly sophisticated cellular models, creating the experimental foundation for a new kind of biological intelligence.
Progress in science is not limited by the number of hypotheses we can generate… we already have many of those. The limiting step is rapid evaluation and verification. In mathematics, software, and engineering, major advances have been enabled by tight feedback loops. Biology has largely lacked that capability.
At @tahoe_ai's scale, those feedback loops become possible. If I hypothesize that drug A acts through pathway B, then C, then D, to produce molecular phenotype E, Tahoe’s data often already contains perturbations that target genes within pathways C and D and should converge on the same phenotype E. Those perturbations become independent lines of evidence that immediately strengthen or weaken my hypothesis. More importantly, when pieces are missing, Rhaister and other models Tahoe has developed leverage the massive scale of data available to infer the missing relationships within seconds, turning observations into a more coherent and continuous biological world model.
To me, that is the real breakthrough. Tara is not an AI copilot that generates hypotheses on demand; it is a true AI scientist, one that turns Tahoe’s experimental scale into biological knowledge that is immediately actionable and generalizable.
English
