> We prove that the time complexity of self-attention is necessarily quadratic in the input length, unless the Strong Exponential Time Hypothesis (SETH) is false[...]
If OpenAI's proof is correct, does that mean self-attention can be implemented in sub-quadratic time? I.e. we could eventually have all the most advanced LLMs running on cheap hardware? To put it more playfully, is ETH being false actually bearish for Nvidia?
[–] _ink_ 2 days ago
[–] caughtinthought 3 days ago
[–] jibalt 3 days ago
What is relevant is that the logic goes the other way: SETH implies quadratic, !SETH implies nothing.
[–] varyherb 3 days ago
[–] margalabargala 3 days ago
[–] refulgentis 3 days ago
But it's also not even wrong, in the Pauli sense.
That makes me wonder if "I see your CS proof, how does this affect LLMs / nvidia stonk price???" being immediately self-described as "playful" ensures criticism looks cranky.
[–] Joker_vD 3 days ago
Turns out, it affected the global politics tremendously.
[–] rokob 3 days ago
[–] dnautics 3 days ago
[–] Xirdus 3 days ago
[–] programjames 3 days ago