These days, they’re usually racks and racks and racks of specialized rackmount servers with all kinds of hardware, hilarious amounts of ram, networked storage, tensor cores, etc stuffed inside, all networked together via fiber optics to run in parallel as one big PC with many CPUs.
- 1 Post
- 396 Comments
Joined 2 years ago
Cake day: July 1st, 2023
You are not logged in. If you use a Fediverse account that is able to follow users, you can follow this user.
yyyyes they do have tensor cores? Where did you get such an absurd idea from?
11·7 days agoBecause a lot of them know that if there was any kind of safety or training requirements, or a psyche or physical test, they would fail it.
5·7 days agoHe explicitly states that no sensitive informarion gets used. If you believe that, then I have
… a bridge to sell you.
Don’t be naive.
I think you’ve misunderstood what I was saying- I don’t have spreadsheets of statistics on requests for LLM AIs vs non-LLM AIs. What I have is exposure to a significant amount of various AI users, each running different kinds of AIs, and me seeing what kind of AI they’re using, and for what purposes, and how well it works or doesn’t.
Generally, LLM-based stuff is really only returning ‘useful’ results for language-based statistical analysis, which NLP handles better, faster, and vastly cheaper. For the rest, they really don’t even seem to be returning useful results- I typically see a LOT of frustration.
I’m not about to give any information that could doxx myself, but the reason I see so much of this is because I’m professionally adjacent to some supercomputers. As you can imagine, those tend to be useful for AI research :P
As I said, anyone with sense.
2517·8 days agoPretty sus how much pro-AI slop is suddenly on lemmy right as GPT5 is about to be launched.
Be less fucking obvious, you dipshits.
Source is the commercial and academic uses I’ve personally seen as an academic-adjacent professional that’s had to deal with this sort of stuff at my job.
Again with the conflation. They clearly mean GPTs and LLMs from the context they provide, they just don’t have another name for it, mostly because people like you like to pretend that AI is shit like chatGPT when it benefits you, and regular machine learning is AI when it benefits you.
And no, GPTs are not needed, nor used, as a base for most of the useful tech, because anyone with any sense in this industry knows that good models and carefully curated training data gets you more accurate, reliable results than large amounts of shit data.
Using chatGPT to recall the word ‘verisimilar’ is an absurd waste of time, energy, and in no way justifies the use of AI.
90% of LLM/GPT use is a waste or could be done with better with another tool, including non-LLM AIs. The remaining 10% are just outright evil.
Those are not GPTs or LLMs. Fuck off with your bullshit trying to conflate the two.
No problem, thanks for being receptive!
Modern day breakers are better about failing safe than they used to be, but… still not an intended use, and all it takes it one bad roll of the dice and everyone gets to have a Bad Day. Hence me viewing them, much like you, as things Not To Mess With Unless Needed. :P
Just so you know, breakers are not on/off switches. Turning them on/off like that is awful for them, they degrade every time they’re tripped. The result of it is, best case, everything on the circuit eventually get destroyed because it’s doing weird shit to the electricity going through it because of a bad connection. Worst case scenario is that it can no longer trip when it should and your entire neighborhood burns down.
While that’s all a good question and thought provoking… unfortunately it’s literally just word of god that if Qui-Gon won, the light would win.
A common problem in Star Wars- the fans have put waaaay more thought and worldbuilding into it than Lucas ever did.
Canonically the actual gamble was qui-gon vs Darth maul. If qui-gon had lived, the universe would have been fine. But he lost, so it was doomed.
The force on star wars is weird, because if you’re strong enough in the force things are basically fated.
Anyway, that’s the long reason why the background music during that fight is called “The Duel of Fates”.
21·1 month agoBy “people like you” I mean people that see this as a good thing. They’re picking and choosing what laws this applies to and what they let slide. This is just the “easy out” that prevents “people like you” from being outraged at the blatant corruption going on.
It wasn’t meant to be insulting.
Oh, don’t pretend that a Republican measure is going to be put under the same scrutiny. This is just an easy excuse so to keep people like you placated with a thin veneer of respectability.
The administration is going to weaponize the FTC anyway, and the supreme Court will back THAT to the hilt.
As for economic effect… That isn’t something the court should be concerned with anyway! Who cares if it’s profitable if it’s illegal!
1·1 month agoI took the lowest end measurement of the worst cohort in the study, actually, and even then they did over ten times the amount of walking you claimed. That’s not a ‘bit’ wrong, that’s off by an entire order or magnitude. And for the record, the entire rest of your post is made up as well- the studies are easily available, the fitness bracelet was given to them, it clearly delineates what bracelets are used, and a wide variety of ages, sexes, occupations, etc etc were studied.
But it was an ingredient, and it’s hilariously tonedeaf to pretend it’s not. Should it be limited where possible? Yes, of course. Hopefully, it’s never needed at all!
But lets not pretend solely peaceful protests have ever worked when they weren’t backed by the threat of violence if the peaceful protestors were ignored. Even MLK Jr. had Malcolm X looming over his shoulder, daring people to try shit… and even then, MLK Jr. still got assassinated for daring improve american citizen’s rights.
As for constantly parroting the Hong Kong protests… you do realize that the Hong Kong protestors lost, right? Why are you using them as a guide for what to do?