Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
raducu
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
20 ms
·
211.
▲
by
raducu
3y ago
> Even if most Western men don’t take multiple wives, men tend to father children with more females than females do with males, a practice called “effective polygamy”. Maybe I'm oversensitive to men's issues, but I feel western
212.
▲
by
raducu
3y ago
> the perversive incentives make them align to just do that. I don't think any therapist actively avoided helping me, but rather at some point they too get comfortable and sort of slack it off like "So what have you been doing&
213.
▲
by
raducu
3y ago
No, I mean borderline personality disorder, bipolar I don't think it goes away with age or therapy.
214.
▲
by
raducu
3y ago
> It's not a fallacy. I feel it is, because it implies it just some statistical trick that's being performed, which is not true at all, imho. I don't know enough about language models, my machine learning knowledge stops t
215.
▲
by
raducu
3y ago
> A therapist trying to play therapist doesn't usually go well either. I've been in therapy for 7-8 years with 2 therapists (well, 4 but one was a couples therapist and 1 only for a few months). Therapy(as well as age, medicati
216.
▲
by
raducu
3y ago
> It guesses words based on previous words. Why can't this fallacy just die already? GPT "guesses" just like ZIP guesses random bits to archive and un-archive files. Except GPT is lossy and IMENSELEY more powerful than
217.
▲
by
raducu
3y ago
And the mathematical amount of genetically unique people that can ever live...
218.
▲
by
raducu
3y ago
> you’re hypothetically depriving a line of offspring of lifetimes they’d otherwise have. I don't want to open a can of worms, but... about 40 milion people are aborted each year...
219.
▲
by
raducu
3y ago
> superconducting islands in a non-conducting matrix. Can somebody who read the article and is knowledgeable in the subject please re-assure me this doesn't mean the material only forms minute super-conducting islands and that we ca
220.
▲
by
raducu
3y ago
That's very interesting! I remember a while ago I read about a room temperature superconductor... but it required enormous pressures.
221.
▲
by
raducu
3y ago
Do you mean superconductors create super capacitors, or how do you want to trap the charge? Or store in coils, as "magnetic energy"(forgot the name for it)
222.
▲
by
raducu
3y ago
I've been completely out of the loop, but please tell me it's USDT?
223.
▲
by
raducu
3y ago
> Not surprising when media hyperfocuses on doom, crime, death and conflict. I used to think the contrary but nowdays I think you're right. A few years ago I unfollowed almost everybody on facebook and my usage time dropped signific
224.
▲
by
raducu
3y ago
But do we know if the amount of temperature drop is reasonable enough for a mass loss, or it's too great and automatically implies supernova?
225.
▲
by
raducu
3y ago
Do you HAVE to tackle it? Is it your idea, to tackle it or is someone else pushing for the problem to be fixed? Is there anyone who knows for sure if it can be done or not? Does it sound reasonable, within the known laws of physics/eng
226.
▲
by
raducu
3y ago
> The first days, weeks and months are often the hardest time when quitting a substance. If this drug helps a smoker not smoke for 3 months, by that time the nicotine addiction has weakened considerably. I love(d) nicotine and it was the
227.
▲
by
raducu
3y ago
> Is this a gradual property that appears slowly as models scale up? Or something that flips beyond a certain size? I wonder if this also applies to biological neural nets. Some animals seem soo close to humans and yet so far. Also if t
228.
▲
by
raducu
3y ago
> solve that last 1% which is exponentially harder than the first 99%. We just don't know that yet! But given how GPT-3 took everybody by surprise and how GPT-4 is so much better, I would not be surprised if we'll be shocked b
229.
▲
by
raducu
4y ago
I get the opposite problem: I often remind ChatGPT that I'm not asking a trick question, because too often when I ask for clarifications about a technology or API, it is too eager to change ita oppinion and halucinate an aswer that it
230.
▲
by
raducu
4y ago
Expedience&censorship and accommodating dictatorships are slippery slopes.
231.
▲
by
raducu
4y ago
Yes, I suspect the AI finds all answers just as easy/hard. Whereas for humans certain programming tasks are orders of magnitude harder than others.
232.
▲
by
raducu
4y ago
But the main reason was slave work that enabled the obscene accumulation of wealth in the hands of so few. Exactly what is happening today with slave wage, acces to AI/automation, discretionary not applying laws(labor laws?) for some p
233.
▲
by
raducu
4y ago
> I don't expect, however, that she will program front ends any time soon Liar :) If she wrote/scored as good as Chat GPT-4 does, you'd be worried she's going to steal your job :). I mean, I know I'm lucky my wi
234.
▲
by
raducu
4y ago
It doesn't really matter how we learn the features that we learn; I've heard numerous times that ML models tend to learn architectures/neurons that biological neurons do; I've seen convincing evidence about image recogni
235.
▲
by
raducu
4y ago
We can't, indeed, but if the past is any prediction -- it will have a few good iterations, better features each time, smaller models, but it will quiet down after a few iterations. Unless Open Ai found some completely novel way of trai
236.
▲
by
raducu
4y ago
> "Fundamentally, LLMs are super-powerful text predictors. Given a prompt, they use machine learning to try and come up with the most likely set of characters that follow the prompt." I've heard this tossed around and in t
237.
▲
by
raducu
4y ago
May I add debugging & maintaining & tuning? I find GPT-4 awesome and certainly it will impact "programming", it's an open question how -- will there be a superclass of GPT enabled programmers that will take the jobs o
238.
▲
by
raducu
4y ago
Anybody who actually read the paper and can share some insights in it instead of spewing opinions and anecdotes?
239.
▲
by
raducu
4y ago
I also expect a high moat, especially regarding training data. But the counter for the high moat would be the atomic bomb -- the soviets were able to build it for a fraction of what it cost the US because the hard parts were leaked to them.
240.
▲
by
raducu
4y ago
> even the good humor and poetry it produces on its own is cherry picked This is so funny in itself, and speaks so much of human expectations :) Yes, the really good examples are cherry picked, but there's a loooot of "good eno
More ›