Links #105

The Wall Street Journal’s front page a week ago. Some powerful white space, that.

Colin Fraser on generative AI being a hammer and nobody really knows yet what is and isn’t a nail. Good essay, worth the read. Seems to me like there are lots of tasks where what you want is determinative logic rather than probabilistic guessing. (Which I think is at least partly why I don’t have much interest in doing AI engineering — I like building determinative things where I can understand how it works. And yes, state machines are totally my jam.)

Ohm, a JavaScript library for building parsers. Looks interesting.

Invisibility shields are kind of real now. Cool. And unsettling. More the latter, I think, given all the malicious ways this could be used.

Jacek Krywko on new developments in active shielding against space radiation. A good-sized chunk of this went over my head (har har), but fascinating nonetheless.