AI designing biological experiments and then getting its hands on the lab equipment is a very different threshold.

Anthropic building a wet lab shows where this is heading: the AI isn't just going to 

 science—it will increasingly participate in the scientific feedback loop, I think.

That could producing incredible breakthroughs. But it also means the old assumption that “the model is only software” is may be becoming obsolete. The most consequential AI may not live in a chatbot window. It may live in a laboratory.

Once AI it can reason → experiment → observe → improve, my question it's  

who controls the lab, the data, the experiments, and the ability to stop the system when something goes wrong?

BooksAndArticles

HealthAndFitness

AI writing code was one thing.

AI designing biological experiments and then getting its hands on the lab equipment is a very different threshold.

Anthropic building a wet lab shows where this is heading: the AI isn't just going to *describe* science—it will increasingly participate in the scientific feedback loop, I think.

That could producing incredible breakthroughs. But it also means the old assumption that “the model is only software” is may be becoming obsolete. The most consequential AI may not live in a chatbot window. It may live in a laboratory.

Once AI it can reason → experiment → observe → improve, my question it's  **who controls the lab, the data, the experiments, and the ability to stop the system when something goes wrong?**