Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
morpheos137
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
5 ms
·
61.
▲
by
morpheos137
4mo ago
yes and it is imposible to verify and evaluate appropriately such information without empiricism. Any empricism LLMs show is stylistic mimicry not a hard coded operational constraint. You can prompt an LLM to test its claims but what it is
62.
▲
by
morpheos137
4mo ago
what is doing the steering is the weights of the words that came before in context. there is no agent or agency. if your problems need median effort and are well represented in shape in the corpus then agents may work well. true inovation i
63.
▲
by
morpheos137
4mo ago
'This is false,' that is an assertion not a proof, and the assertion rests on some kind of asbolutizing fallacy where if x =! y then y =! y(x). I agree cognitive states do not require language. However to say language does not mod
64.
▲
by
morpheos137
4mo ago
ok however i would say extrapolating the current data set is not a way to exceed the the human envelope. it is unclear to me the human evelope has been demonstrated as a convex hull or how transformers could find points outside it. in other
65.
▲
by
morpheos137
4mo ago
again, try thinking critically it is not merely an interpolator means it can interpolate on many dimensions. it does not follow that greater than human capability results from doing so. explain to me how a statistical function approximator
66.
▲
by
morpheos137
4mo ago
allow me to resolve the confusion: ai is a model of language. language is a model of human cognitive state. to the extent language maps to cognitive state accurately and ai processes language fluently ai models understanding. whether this i
67.
▲
by
morpheos137
4mo ago
ai can abitrarily closely fit the human corpus. why people expect it to magically achieve superhuman qualities is beyond me. we got a very good statistical interpolator. how do you go from there to superhuman when training is on the human c
68.
▲
by
morpheos137
4mo ago
kind of funny tokens don't prompt and steer themselves. it almost as if the value still lies with the human holding the tool.
69.
▲
by
morpheos137
4mo ago
kinda meta lol but my incredible depression was performatively cynical.
70.
▲
by
morpheos137
4mo ago
its incredibly depressing ostensibly intelligent people get depressed about others having different points of view or set up fallacies of the excluded middle / xor fallacies where not warranted.
71.
▲
by
morpheos137
4mo ago
people need to define what "understand" means before they argue about it. example, I as human do not understand what: "The first warning was about scale itself. Bender and Gebru argued that training ever-larger models on ever
72.
▲
by
morpheos137
4mo ago
the real interesting way to address the question of token effectiveness would be internal alpha vs beta testing and measuringing marginal revenue generated by similar teams using ai and at different usage levels. right now $1500 a month is
73.
▲
by
morpheos137
4mo ago
AI is the interpolation of the human corpus. Is it suprising AI recombines sucessfully where human attention has not explored all plausible solutions? N,o not especially. The key fallacy is that AI is other than human. This is really no dif
74.
▲
by
morpheos137
4mo ago
yeah i was replying to the comment that selling intelligence on private citizens to police would be a revenue driver. not that it couldn't be done, just that it would not make significant money. the median critical thinking level on HN
75.
▲
by
morpheos137
5mo ago
yeah the parent was a rather dumb comment. police don't buy intelligence on the population en mass. how could they afford it snd how could they use it without probable cause? if you are already a target of an investigation they can get
76.
▲
by
morpheos137
5mo ago
have it write python porogram to generate the svgs. then use the program. circuit diagrams are rrlatively thin corpus but it knows how ciruits work sufficently to write a program.
77.
▲
by
morpheos137
5mo ago
the fallacy is not recognizing ai is interpolating the human body of work. all the people decrying ai slop: this is our sythesis. ai is not a separate thing from human work. whwt the author is really saying is "I am tired of individual
78.
▲
by
morpheos137
5mo ago
I'm interested in the program he would have used to typeset his presentation in the 1960s. Also not at all an attack but he seems to be either nervous public speaker or to have a well mananged speech impediment. Just an observation not
79.
▲
by
morpheos137
5mo ago
https://news.ycombinator.com/item?id=48225426 there is a difference between a stunt and a viable product. diverless cars and agi are the fusion of Silicon Valley.
80.
▲
by
morpheos137
5mo ago
emailing
81.
▲
by
morpheos137
5mo ago
there is little evidence for this prediction.
82.
▲
by
morpheos137
5mo ago
it seems obvious to me the ability to push code to public repositories shluld be tied to real human identity.
83.
▲
by
morpheos137
5mo ago
That is a dubious asertion. At high enough resolution the model converges on the modeled.
84.
▲
by
morpheos137
5mo ago
Mention "consciousness" get a slew of incoherent responses. There is no "hard problem." the burden of proof has always been on those who claim there is. whats it like is what it is what it is is what it is like. the ques
85.
▲
Dealing with a Fake a World
2 points
by
morpheos137
5mo ago
|
2 comments
86.
▲
by
morpheos137
5mo ago
The Yudkowskian 'Bayesians' are a scifi cult. Intelligence as it exists in nature is neither arteficial nor general nor 'super.' All human technology is natural. It certainly is not supernatural.
87.
▲
by
morpheos137
5mo ago
OP here. I solved all my problems but the point is in OpenBSD and in my experience linux these would not be problems in a default install. Any now that it is meeting my needs happy with netbsd.
88.
▲
Gemini Android App User Hostile Behavior
1 points
by
morpheos137
5mo ago
|
0 comments
89.
▲
by
morpheos137
5mo ago
i am not understanding why could'nt a deterministic dictionary program do it?
90.
▲
by
morpheos137
5mo ago
AI is completion generation. if appropriately constrained, prompted, and monitored it can generate completions at par with or even exceeding the best human material. the question is is the control cheaper than a human expert given it requir
More ›