@rrwilliams @Timeroot Geez, AI companies rediscovering and reinventing peer review

@Vinod_MIT this is a really impractical ask. just imagine if you were tasked everyday to write a post here on everything you thought of and why it did not work. would you do it?

@ylecun @huskydogewoof Appreciate the historical perspective. Perhaps you could also explain *why* it doesn't scale. This would make it more educational.

@pstAsiatech ok the chinese distill and nvidia can only distill from them

@dylan522p @negligible_cap When I shouted "Fire, Fire!" in the crowded theater, I was talking to a colleague who deserved to be fired. Anyone who panicked simply misunderstood the context, so the fault lies entirely with them, not me. Nada. Zip. I am the golden boy who can do nothing wrong.

omfg - are people seriously having this discussion about consciousness? If it doesn't feel, which it doesn't because it doesn't have sensory receptors like smell, touch and taste, it can't have consciousness. It is as simple as that. Thank you.

there is no ai battles only narcissistic pundits who want max spotlight on them.

@ylecun @Noahpinion @ylecun none of what are you describing are not necessary for intelligence. People who are deaf, mute, blind or incapable of moving have achieved plenty of great things.

superintelligence! ammerrricaa.

@dinodaizovi This is NOT about model weight exfiltration.

@Vinod_MIT @rohinmshah Now that you have written this paper the agents have training data to exactly do this in a later version! Talk about inventing the future.

@siliconcodesign @OpenAI @hotchipsorg actually the opposite! for the accelerator work they did they had a really large team! And they used the compilcated IP pieces --- the core is simple!

would hate working at openai and having to sit thru genai generated slides with velocity --- why does everyone speak like this

@lauriewired If the specs hold up for real benchmarks it will blow up so many arm server plays esp. in the high reliability market. Just because it has failed in the past doesn't mean it will this time.

While i was on my run today I casually asked Claude to solve global warming because it was so hot outside. I said "make no mistakes". Eagerly awaiting the moment when I can announce the result here!

@random_walker My man, your analysis is hollow; your writing shallow; and your mind fallow.

@matthew_d_green really man? mceliece is shady? that's the only scheme where you can even get a tiny bit of a tiny ounce of intuition on why it *may* work. tell me honestly you understand lattice crypto?

#mathocalypse now

@NaveenGRao there are many ways to build useless "computers". after 80+ years of experimentation we have come up with few economically feasible ways to build them.

@philvenables @ZackKorman I think what is being said is that certain people who are traditionally part of the community -- such as independent contributors/researchers, small businesses and perhaps even university researchers -- are not being included.