High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / belief

Published · transcript-backed

Alessio Fanelli: belief

31 Dec 2024 Latent Space Latent.Space 2024 Year in Review

“I think that's like one of the hardest things about NeurIPS. It's like the long tail.”

— Alessio Fanelli

Source trail

Everything needed to verify it.

Speaker
Alessio Fanelli
Attribution
Verified speaker
Claim type
belief
Recorded
31 Dec 2024
Publisher
Latent Space

Transcript context

…someone, if the only message I give you is meet me in New York and you're not aware. Or when you would probably meet me at Grand Central Station. That is the Grand Central Station is a shelling point. And it's probably somewhere, somewhere during the day. That is the shelling point of New York is Grand Central. To that extent, shelling points for steganography are things like the, the, the common decoding methods that we talked about. It will be interesting at some point in the future when we are worried about alignment. It is not interesting today, but it's interesting that DeepMind is already thinking about this. I think that's like one of the hardest things about NeurIPS. It's like the long tail. I found a pricing guy. I'm going to feature him on the podcast. Basically, this guy from NVIDIA worked out the optimal pricing for language models. It's basically an econometrics paper at NeurIPS, where everyone else is talking about GPUs. And the guy with the GPUs is…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.

Search evidence