There was a long time during which NNs were considered a dead end. Pretty good that Geoff Hinton, Yann Le Cun, and others kept plugging at it.

F'ck I am so tempted now.

Assuming that being immortal makes civilization stagnant and/or happy seems like a tall assumption. 🤔

At low enough rate, long mortgages are asset not debt. Anyone who was smart took as large as possible 30y before rates went up (~3% rate) and would have taken 50y if it was possible.

No, it's not just rent. Your property price rises but mortgage does not.

I dare the Dare Market founder to jump out of the plane without a parachute. For publicity reasons, obviously.

I write my papers in a mountain temple while sitting in lotus position with an ancient calligraphy brush.

If you subscribe to Paramount+, you can also chat with South Park Satan.

"Move over, I am going to fix the broken leg and stop arterial bleeding using my homeopathic remedies!"

I submitted a piece of performative AIrt to NeurIPS, but for some reason it was rejected with extreme prejudice.

Making America great waves again

Kindle version will be available?

bsky.app/profile/jurg...

Loading data... (loading data message will be improved in two years)

I am a reviewer and AC for top AI conferences and I've seen accusations of papers and reviews being LLM-generated, but I never ran LLM check. We have review process, so it's different from your situation. How to stop deliberate scammers is an issue. I don't think even with human review it's solvable

I think using ChatGPT or other LLM detector is risky. Too many false or semi-false positives. Author may have used LLM to fix language or style. Is that bad? Does LLM detector distinguish "LLM content generated" from "LLM style generated"? I don't think it does or even can.

Who let the skunk into the house? 🐯