804 karma · joined May 26, 2018
It’s definitely not “too hard” for a professional. If someone can’t even do this much, their reporting is not trustable in the first place.
Scaling diagrams also showed no sign of plateauing AFAIK.
Wrong. Various regularization schemes are used in AI models which essentially introduce “flaws” and noise into the process. The flaws are more optimal than what the brain does, but they are there.
I have seen this line repeated many times, but I never saw it actually explained. A lookup table is dumb and easy to understand/interpret. Deep models are not that. They are also not a linear interpolation of … something. What exactly is the claim being made here? Yes, deep models don’t generalize too well on ood data. How does this make them a “very optimized” lookup table?
I don't believe so. In general, assholes who get their money from the government, create their own fake news which is even faker than what is published in the state media. Which makes sense, as the state media optimizes for everyone watching it, not just the bootlickers.
> regular VPN protocols
Naive, Hysteria, Trojan, VLESS, etc. seem to work. Of course, foreign servers are, at times like this, only accessible in datacenters, which means the protocol would be irrelevant.