|
sony doing exactly what i wanted apple to do if they were going to mess around with cars anyway, just get honda to build the car and focus on making it not insufferable. not that i necessarily trust sony to succeed in that, but can't be much worse than a car company.
|
#
¿
Jan 5, 2023 12:28
|
|
|
|
| # ¿ May 10, 2024 08:41 |
|
|
haveblue posted:it does seem weirder to render yourself blind in a public space drones are awesome for this sort of thing though.
|
|
#
¿
Jan 8, 2023 22:18
|
|
|
they've been working on this for like a decade, semi-cancelling it twice, it'd be a miracle if it comes out good.
|
|
#
¿
Jan 12, 2023 10:23
|
|
|
CommieGIR posted:Sure it's easy to pass an exam when you can just copy/paste from your massive amount of training data yeah, i think some are taking this as a slam against the exam (the subject area is largely ridiculous of course, but i doubt the exam is substantially different than many in other more relevant areas), but really if a student got a year to write the exam with full access to the internet and literature, i imagine just about anyone would pass it. and purely inference-time and information it has access to chatgpt is not that badly understood in that way. (but i consider chatgpt an impressive achievement, great potential for a leap forward in lots of areas, just almost all of them a touch more mundane than what the internet imagines)
|
|
#
¿
Jan 25, 2023 12:41
|
|
|
90% of all sd derived models, no matter what the headline purpose of them is, have just been finetuned with a bunch of weird internet porn.
|
|
#
¿
Feb 1, 2023 14:09
|
|
|
the problem we wanted to solve as natural language processing, while it is somewhat ill defined it is way more solved than anyone expected. the chatbot stuff is weird tech demos getting interpreted in even weirder ways by a mostly clueless public and an entirely and willfully clueless press.
|
|
#
¿
Feb 24, 2023 16:43
|
|
|
infernal machines posted:turns out testing responses to novel situations is hard not the least because you constantly have to come up with novel tests models using training sets predating superglue have done great on superglue though. the actual nlp progress is not imaginary. the conversational stuff is entirely imaginary though, both in being a nonsense application and in results being extremely frail.
|
|
#
¿
Feb 24, 2023 18:44
|
|
|
i will presume that is chatgpt doing a particularly poor job of making that point though e: to be clear, literally, as the chatgpt response i get is similar but better. Cybernetic Vermin fucked around with this message at 19:00 on Feb 24, 2023 |
|
#
¿
Feb 24, 2023 18:54
|
|
|
infernal machines posted:not being a genius mind at a top tech firm, this would seem to me like it's a poor fit for a product that relies on conversational interaction, but maybe that's why i'm not working at microsoft or google yeah, i do think it is pr nonsense atm. i think google is right to be a touch worried though, as one thing they might be able to do is make more robust search engines. i.e. google killed off structured search engines, with boolean operators and whatnot, but a llm might actually be able to robustly turn stuff like "what's a nice non-vegetarian main dish that is nonetheless gluten- and dairy-free" into "recipe and (chicken or steak or ... or cod) and not vegan and not vegetarian [etc.]", which would make a pretty good tool for real people. not fancy enough for pr and demos, but useful enough to make another search engine a contender.
|
|
#
¿
Feb 24, 2023 19:07
|
|
|
it absolutely can translate what you say into some other system though; the dumb bing bot does indeed do bing searches. sagebrush is both right and wrong, in that people are not a good metric for intelligence, but otoh people by and large do not seem to be misunderstanding chatgpt *that* deeply (outside a handful of people, like nyt headline writers and that one google engineer).
|
|
#
¿
Feb 24, 2023 22:31
|
|
|
also lets move to the chatgpt thread please, i love the cyberpunk thread for *not* being about extremely shallow current tech garbage.
|
|
#
¿
Feb 24, 2023 22:33
|
|
|
CommieGIR posted:Not that I disagree about moving it but cyberpunk is all about shallow tech garbage! in regards to technical depth maybe, but if there's one thing that defines cyberpunk it is deep societal commentary. and while gpt stuff might get there, it for sure isn't happening, or is obviously about to happen, now.
|
|
#
¿
Feb 24, 2023 22:56
|
|
|
finally a gpt model for english
|
|
#
¿
Mar 17, 2023 13:50
|
|
|
i do a lot of care-mad-posting about how ridiculous hype in ai is causing all kinds of harm, but here at least we have something feasible, creating a machine that operates at the same cognitive level as diane feinstein is easily within our technological reach.
|
|
#
¿
May 22, 2023 13:41
|
|
|
haveblue posted:if you read the article it's very clearly a joke feel like the link contains all the information necessary here Cybernetic Vermin fucked around with this message at 10:41 on May 23, 2023 |
|
#
¿
May 22, 2023 22:32
|
|
|
Beeftweeter posted:lol yeah thats why the dichotomy of "hey this thing is pretty good, i'll use it" or "gently caress this thing is too good, it took my job" was so amusing ironically i think the wga strike is entirely correct to make a big deal about ai scripts anyway, not because i think it is a feasible thing to do, but because of hype making people think that it is feasible there's an opening for studios to hire non-union writers as "ai editors" pretending that the thing is written by ai despite it being all human input all over.
|
|
#
¿
May 24, 2023 08:18
|
|
|
either way, just add some good tunes to the queue imo, free music buddy
|
|
#
¿
Aug 10, 2023 15:52
|
|
|
yeah, i love the fable games, but the things that are good about them are almost all things entirely unrelated to the things molyneux talked about. and that's sort of the thing really, he talks a lot of garbage but the hit-rate of games he's involved in is pretty drat high. can only hope that this latest thing launches without all the crypto nonsense and it turns out to just have been molyneux being molyneux.
|
|
#
¿
Oct 3, 2023 11:00
|
|
|
ymgve posted:i mean if its in addition to the real windows its fine nah, i think that's dumb. in the unfortunate circumstance where you have a room with a need of more "window", get some suitable natural light panels, and a lot of plants. i mean, maybe some screens would also liven things up, but there's no way you will ever fool your reptile brain into that 2d vista being real, so might as well then put up some art you enjoy.
|
|
#
¿
Oct 16, 2023 20:38
|
|
|
NoneMoreNegative posted:lol when someone launches a forums bot and the first post lands in FYAD and is immediately probed. that'd be a lot funnier than the reality though
|
|
#
¿
Nov 6, 2023 22:05
|
|
|
|
| # ¿ May 10, 2024 08:41 |
|
|
haveblue posted:I don’t see how the second thing will be any less large, heavy, complicated, or expensive than a car to begin with if they worked i could see them being a rental thing in some circumstances. low speeds around town to haul large purchases home etc. i don't see them working without insanely dangerous failure modes though.
|
|
#
¿
Dec 29, 2023 07:52
|
|