Outside of academia there are only a few niche industries still hiring. Mag7 is drying up. The semi-private research institutes want seniors with grants or customers in pocket, not fresh phds with no connections.
Probably only getting worse in the near term.
There are a few specific applications that are still good. Medical imaging seems okay for now. Advanced signal processing is still a viable route. Consumer robotics, possibly.
Short-term: you better love it. Long-term: just focus on being one of the ones that ensures humans retain their ability to understand math. Do a YouTube channel on the side where you make math legible to the masses in a fun, engaging way. Use the cash from that to fund your research/genuine passions.
Whatever you do: please don't submit to the machine or throw away your genuine curiosity as a sacrifice on the Altar of Commerce.
It's also relevant for non-AI reasons. The upshot is: Do it if you want to. You're at an age where you're supposed to be mostly learning, but over time, start transitioning into whyever you did this (which doesn't have to mean staying in academia, and in fact can mean teaching or popularisation or consulting or...).
Oh, no - that's actually now.
LLMs have yet to show that they can meaningfully make such helpful abstractions. Not saying that it can't be done, but I wouldn't write such doomer posts just as yet.
Unsolved Erdosh problems were touted as a cheap way to generate new perspectives. outcome has been slightly disappointing. overlooked frameworks were all that has been needed so far. not new ones. Could change as LLMs are pointed at other kinds of inexplicables
The students who have been alarmed outside his office should take a course in marketing and journalistic ethics to assess the situation more rationally and figure out that the problem might sit in the office.
It is interesting by the way that Google does use Lean in a loop to refine proofs. This was called heresy by AI boosters earlier who said that Lean was never used in proofs.
IIRC he explicitly stated that he doesn't.
What did Kurzweil or Yudkowsky predict that actually came to pass?
I assign this Scott no points for bringing up Penrose as a straw man. That’s a very old canard.
