xdavidliu

i don't understand what the second paragraph is saying.

less intense than in countries near equator, not less intense than winter

only if it is not littered with unsafe blocks

pull requests are welcome

adding to the irony is the fact that the mechanism used by the NN architecture in LLMs is also called attention

its not that they did not get your pun, its that they think building rockets is hard, and i tend to agree

this comment is nonsensical

not a bug; the post we are currently on is the one that apparently used to be negative.

oh the irony

framing 4 gb of data per user as 4 gb is even more of a stretch

I'm of course not talking about the arithmetic, but rather understanding the semantics of ratios, sets, and subsets, and inferring statistical significance.

I think the CEO was more talking in the line of Bill Russell or Maximus from Gladiator, not final-year Haslem

this is the type of math they should be teaching in high school, not trigonometry and calculus (which should be electives)

humans are probably the least cheap thing you can have in this context

404 not found

"delve removed from y combinator" removed from y combinator

kinda reminds me of the plot of Sphere, where Samuel L Jackson is reading 20,000 leagues under the sea and is thinking of giant squids.

that would be the exact opposite of modern

fair point

good question, I'm not sure. Maybe check out the new Eliezer Yudkowsky book? He definity talks about something akin to "post-GPT era" on there.

I should have clarified what I meant. The training data includes roughly speaking the entire internet. Open source code is probably a large fraction of the code in the data, but it is a tiny fraction of the total data, which is mostly non-code.

My point was that the hypothetical of "not contributing to any open source code" to the extent that LLMs had no code to train on, would not have made as big of an impact as that person thought, since a very large majority of the internet is text, not code.

for a book that surveys pretty much all of it, see "Empire of AI" by Karen Hao

seems to me when this kind of stuff happens, there's usually something else completely unrelated, and your comment was simply the first one they happened to have latched onto. surely by itself it is not enough to elicit that kind of reaction

open source code is a miniscule fraction of the training data

i think your comment actually mostly makes sense, except the part about neural network guys needing to familiarize with Chomsky, which is not the case at all

yeah, your parent comment appears to be just nonsensical name-dropping. happens a lot here. A different type of comment is the "X is just Y" comment that is kind of annoying, like "all of AI is just curve fitting", which the commenter wants readers to think is some kind of profound insight.

when Elon bought twitter, I incorrectly assumed that this was the reason. (it may still have been the intended reason, but it didnt seem to play out that way)

is there a more accepted connotation of the lone word "computation" that means something different from "theory of computation" (in the sense of turing machines, computability, decidability, complexity classes, Sipser) etc?

One time in around 2008 when I was in undergrad, I got facebook requested by a guy named David Liu (same name as me) who was going to a school many thousands of miles away, and I noticed that he had facebooked about 20-30 other David Liu's.

About 3-4 years later, I was in grad school, I meet this guy in person (he happened to be going to the same grad school), I recognize his name and face, and let him know that he facebooked me back then, and he got a chuckle out of it.

I got the same impression as well. I think I've become so cynical to these kinds of things that whenever I see this kind of thing, I immediately assume bad faith / woo and just move on to the next article to read.

i watched the lecture series during the pandemic and commented on many of the youtube videos. in at least one instance, a library function is used on the board that is not compatible with the current function signature in mit scheme.