LESSWRONG
LW

All of TheSupremeAI's Comments + Replies

I would also gamble 200:1 odds for up to $5000. I have a strong conviction that all UFOs are prosaic in nature and have held this conviction for longer than a decade at this point.

1RatsWrongAboutUAP2y

Acceptable offer, reach out with payment details

NVIDIA and Microsoft releases 530B parameter transformer model, Megatron-Turing NLG

TheSupremeAI3y10

While far from claiming numeracy, the model seems to go beyond only memorization for arithmetic.

Pretty optimistic that AGI is going to be achieved by the end of the decade

3Capybasilisk3y

Leo Gao thinks it might be possible to ride language models all the way to AGI (or something reasonably close): https://bmk.sh/2021/06/02/Thoughts-on-the-Alignment-Implications-of-Scaling-Language-Models/

2Quintin Pope3y

I don't think that's likely. Rates of increase for model size are slowing. Also, the scaling power-laws for performance and parameter count we've seen so far suggest future progress is likely ~linear and fairly slow.

1John Steidley3y

I can't tell what you mean. Can you elaborate?

DeepMind: Generally capable agents emerge from open-ended play

TheSupremeAI4y50

And more compute of course. As always.