Discover more from Dear Future
A.I. revolution incoming
Except apparently in chemistry where it might actually be beneficial
Is it hyperbole to compare the push by tech companies (and governments) to gain supremacy in A.I. technology with the Manhattan project and the invention of the atomic bomb? Before listening to this Ezra Klein interview with Kelsey Piper—about the increasing pace of A.I. development and how it’s changing the world—I would’ve said that the main concern was job loss for knowledge workers and creatives.
Now I’m not so sure. In the next decade, A.I. is very likely to be replacing knowledge workers — up to and including writers, lawyers, and investment bankers — but that’s not the biggest concern. The threat is likely not what you or I imagine, but that’s part of the issue: namely, that development is moving so fast down the capitalism-greased paths to make money for the A.I. development companies that society can’t adjust fast enough for the unexpected.
The same thing happened before, in the rush to create the atomic bomb. The fictional narrative driving the Manhattan project was that we needed to get there before the Nazis. Turns out they weren’t working on it at all.
Listen or read:
Meanwhile in chemistry…
Almost every chemist agrees that generative A.I. systems can massively accelerate research into protein folding, retrosynthesis, and novel drug identification. But that’s only possible if there’s enough data shared to effectively train the systems.
Today’s editorial in Nature calls for more open sharing of chemistry data, including negative results.