> Maybe Sabine Hossenfelder tries to use them for things they can't do well and she's not aware that they work for other use cases.
Yeah. That was my first thought. There’s probably orders of magnitude more training data for software engineering than for theoretical physics (her field). Also, how much of software engineering is truly novel? Probably someone else has already come up with a decent solution to your problem, it’s “just” a matter of finding it.
Yeah. That was my first thought. There’s probably orders of magnitude more training data for software engineering than for theoretical physics (her field). Also, how much of software engineering is truly novel? Probably someone else has already come up with a decent solution to your problem, it’s “just” a matter of finding it.