calf

16 ms·

271.
by calf 1y ago
You said that classic LLMs have no inductive bias for causality. So I am simply asking if any computer scientist has actually proved that. Otherwise it is just a fancy way of saying "LLMs can't reason, they are just stochastic par
272.
by calf 1y ago
But then that sounds more like that person explained it wrong. They didn't explain why it is necessary to reduce to GRAPHCUT, it seems to me to beg the question. We should not assume this is true based on some vague anthropomorphic app
273.
by calf 1y ago
How do we prove a trained LLM has no inductive bias for space, causality, etc.? We can't assume this is true by construction, can we?
274.
by calf 1y ago
Above commenter also asserts "highly developed research but no public literature" shrug ...
275.
by calf 1y ago
[flagged]
276.
by calf 1y ago
[flagged]
277.
by calf 1y ago
"Evidence-based" is a really problematic term when it is used to protect bureaucracies and medical managerialism, rather than actually interact with scientific processes in an ethical way. Their anecdote is actually a good example
278.
by calf 1y ago
What valid AI criticisms? Most criticisms of AI are not very deep nor founded in complexity theoretic arguments, whereas Yann LeCun himself gave an excellent 1 slide explanation of the limits of LLMs. Most AI criticisms are low quality argu
279.
by calf 1y ago
The extended Church-Turing thesis is specifically about the relationship between theoretical TMs and the physical universe. So these paper authors are just begging the question—they disagree with the thesis. But as I say in a) and b) above,
280.
by calf 1y ago
Their argument rests on computation being a theory ("simulation") while agency/cognition being real ("processes"). Put that way, I don't buy the distinction. Specifically, my reactions are: a) Defining agency i
281.
by calf 1y ago
Hinton is a domain expert in AI, case in point Scott Aaronson cites AI revolutionizing protein-folding research, putting many protein-folding academics out of jobs. I'd argue that AI experts and CS theorists are in a better position to
282.
by calf 1y ago
Was Knuth talking about deployed software when he wrote that?
283.
by calf 1y ago
Their TV app changed grid layout of playlists to list that scrolls down so slowly, this makes my 100+ video playlist useless! Argh.
284.
by calf 1y ago
Do non-AI coding assistants provide value?
285.
by calf 1y ago
If I were to fathom a guess it would be (the argument that) that LLMs have been a science experiment upon humanity, that many of us haven't consented to. Imagine if OpenAI instead crawled the Earth for shedded human hair and skin cell
286.
by calf 1y ago
That carryover is because Turing machines and Lambda calculi are equivalent models of computation, and not because "busy beaver" is some special concept. It's a function whose definition is about computational phenomena, so t
287.
by calf 1y ago
What precisely is meant by simulating the laws of physics? There's a common understanding pitfall there, in that a lot of lay people quickly end up talking about "is the universe a computer" or "infinite Turing machines
288.
by calf 1y ago
AI for writing or research is useful like a dice roll. Terence Tao famously showed how talking to an LLM gave him an idea/approach to a proof that he hadn't immediately thought of (but probably he would have considered it eventual
289.
by calf 1y ago
Yann LeCun gave a recent talk, at 10:30 he gives a good explanation that doesn't appeal to "what LLMs are trying to do on the inside": LeCun just observes that next-token prediction is exponentially diverging, since every nex
290.
by calf 2y ago
What I think lends strength to Chomsky's theory is that is virtually, informally a corollary of computational complexity theory (grammars, P = NP, and related ideas) which was a direct consequence of Alan Turing. So for the camp of lin
291.
by calf 2y ago
I don't get it, so dire wolves were only 20 gene changes from gray wolves? Not thousands of tiny,crucial changes all over their respective genetic codes?
292.
by calf 2y ago
What bothers me is the ageist assumption that "full-adults", say, boomers, are somehow more educated, less indoctrinated, or less prejudiced than young adults
293.
by calf 2y ago
Tribalism is just really bad pop-sociology, by people who can't be arsed to read and do their homework on a vast subject matter.
294.
by calf 2y ago
Tribalism is just bad sociology, that's where the nuance is missing.
295.
by calf 2y ago
Then the center-left Democrat progressives need to stop discussing the term "free speech absolutism" because it conceptually muddies the water. Leftists believe in free speech, freedom of inquiry, freedom of the press, even "
296.
by calf 2y ago
I was going comment how much I enjoy using Blender's UI and not GIMP, though both are OSS.
297.
by calf 2y ago
I don't see how this is a proper reply to the prior comment; which Chinese lecturer has a Nobel prize in AI?
298.
by calf 2y ago
Those four parts sound like one unified, cognitive algorithm -- having an ontology of the problem by breaking it into subgoals; checking your work properly; thinking backwards to debug a mistake and retrying; and thinking ahead and reasonin
299.
by calf 2y ago
I'm slightly terrified I'm mot consciously trying any of these metacognitive things to "self improve" myself.
300.
by calf 2y ago
Maybe Zelensky had a very thin path out after Vance's provocations, but even the language accents are partly to blame for miscommunication/crosstalk, so maybe for Zelensky it was a Kobayashi Maru situation.

More ›