Dan GoodmanMar 9
If you take a paper and get an LLM to review it, then get the LLM to rewrite the paper and write a reply to the reviewer, and then repeat, what happens? Convergence to something better? Cycles of arbitrary change that never converge? Descent into meaningless drivel?
Show threadAlbert CardonaMar 9

@neuralreckoning

Loss of what the point of the paper was.

Show threadDan Goodman
@albertcardona just like real peer review then? 😉
Mar 9 at 11:25pmWeb
Show threadAlbert CardonaMar 9

@neuralreckoning

Snark aside, I find peer review helpful, to a point. Human peer review, which is about ideas, vision, concepts, and suffiency of methods and data, not about pretty turns of sentence.

Show threadDan GoodmanMar 10
@albertcardona you're right I'm sorry I just couldn't help myself. The goal was wide open.
Show threadAlbert CardonaMar 10

@neuralreckoning

How's PCI neuro going? I'm interested in the model; a pity though that there are reject decisions, unnecessarily so in my opinion.

I see @brembs and Koen Vervaeke in the board as well.