make3

37 ms·

601.
by make3 4y ago
It's funny that something that is now a super basic approach that has been used literally everywhere for 20 years is presented as some fancy Nintendo performance hack
602.
by make3 4y ago
arguably your brain also learns a representation of an algorithm too
603.
by make3 4y ago
it makes shit up still
604.
by make3 4y ago
they haven't given any sign that they will limit the access. They have given signs that they are capitalists & are ready to do a lot to make money, like not putting a list of authors on the GPT4 paper & not write anything about
605.
by make3 4y ago
it definitely learns algorithms
606.
by make3 4y ago
the problem is that if you steal the weights then you can serve your own gpt4, and it's very hard to prove that what you're serving is actually gpt4. (or you could just start using it without paying ofc)
607.
by make3 4y ago
doing that would make it less likely that they would publish a model next time
608.
by make3 4y ago
fired usually means, with cause. laid off is for when a lot of people are getting cut at the same time because upper management asked to do big cuts
609.
by make3 4y ago
you can rent a vm on aws to run it
610.
by make3 4y ago
it's really easy to get access to the weights, I got access to them as a master's student with just a single very minor publication in the domain, and basically all my colleagues who applied did too. there is no Aaron Schwartz thi
611.
by make3 4y ago
Michael Reeves is one of the best "wide public" tech youtubers
612.
by make3 4y ago
stack overflow* the economy
613.
by make3 4y ago
but that's just because tech crashed
614.
by make3 4y ago
I feel like one of the only ways Youtube can be profitable is by having integration and being a first class citizen Google cloud's infrastructure & tech.
615.
by make3 4y ago
do they do instruction fine-tuning
616.
by make3 4y ago
your tool would not be thought of as a reasonable source of information by anyone, including the court, who are not mindless automatons but understand context & your disclaimer is a lot more visible than it is on bing etc.
617.
by make3 4y ago
you're not presenting them to the public when they're asking something relating to that fact though
618.
by make3 4y ago
you're holding their identity hostage like that? I would just not talk to you
619.
by make3 4y ago
he never said it was hard, his point is that it's unintuitive & a lot of people don't know or don't remember
620.
by make3 4y ago
i think you're confusing slack and discord, everything is in threads in slack
621.
by make3 4y ago
Well they had a giant conversational size bigger than GPT3. It was not as safe or good at following instructions as ChatGPT. But it was as flowing as it is. So maybe it's an exaggeration (this word has two gs in english!?) to say that
622.
by make3 4y ago
as someone who did nlp research at Google, Google had a ChatGPT level chatbot 3 years ago and did nothing with it
623.
by make3 4y ago
what's your point? that it's not as good as a human? I don't think anyone is saying that. people are saying it's impressive, which it is, seeing how quickly the tech grew in ability
624.
by make3 4y ago
or they could be used to censor all content too by doing simple classification
625.
by make3 4y ago
the products that both attacks the bacterias & are perfectly safe are probably very very hard to find, right. the product needs to be safe to flood your whole system with, because that's how pills work, you eat it and it diffuses e
626.
by make3 4y ago
as much shit as we can give them for something as trivial as their name, their contributions to AI with GPT-1 (which beat all benchmarks & which led to BERT & was basically the "real" BERT itself), GPT-2, which was really
627.
by make3 4y ago
compulsory vote in the US & Canada would be interesting. It also has no chance of happening as the right in the US is focusing so much energy making it hard for minorities to vote because they know it would fuck them
628.
by make3 4y ago
I don't understand the end of your sentence
629.
by make3 4y ago
talking about left wing discourse as a single unified entity is insane, when it's millions or billions of people with different beliefs, formations, etc. Like in everything else, some discourse will be productive and some will be misgu
630.
by make3 4y ago
if there is not enough data then all we can do is increase model size, which supports my original point

More ›