this post was submitted on 10 Dec 2024
109 points (100.0% liked)

TechTakes

1491 readers
47 users here now

Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

founded 2 years ago
MODERATORS
 

I can't wait for the spectacular implosion

top 50 comments
sorted by: hot top controversial new old
[–] Th4tGuyII@fedia.io 30 points 2 weeks ago (3 children)

I think AI tools have (and will have) their uses, but AI as a whole has been hyped up so much for so long now that the bubble is bound to burst sooner or later.

And when it does, hopefully we can go back to not having AI shoved into every fucking thing imaginable...

No I don't want an AI on my phone or computer trawling through all my data for you, just to give me some handy search feature I almost certainly won't use.

[–] GregorGizeh@lemmy.zip 19 points 2 weeks ago (2 children)

I might want an ai on my computer, but only if it is a local, open source model that does not report any kind of data to outside parties in any way.

[–] Th4tGuyII@fedia.io 7 points 2 weeks ago

If that were the case, and it was something I chose, I certainly wouldn't mind it anywhere near as much - but the ones being forced upon you by every tech company alive right now are none of those things, and are all data harvesters disguised as utilities.

[–] db0@lemmy.dbzer0.com 15 points 2 weeks ago (1 children)

I expect a creative destruction, like what happened with the dotcom bubble. A ton of GenAI companies will go bust and the market will be flooded with cheap GPUs and other AI hw which will be snapped on the cheap, and enthusiasts and researches will use them to make actually useful stuff.

[–] dgerard@awful.systems 11 points 2 weeks ago* (last edited 2 weeks ago) (1 children)

these are compute GPUs that don't even have graphics ports

[–] db0@lemmy.dbzer0.com 15 points 2 weeks ago (2 children)

Yes, my point is that the compute from those chips can still be used. Maybe on actually useful machine learning tools that will be developed latter, or some other technology which might make use of parallel computing like this.

I'm waiting on the a100 fire sale next year

[–] JackRiddle@sh.itjust.works 6 points 2 weeks ago

I know of at least one company that uses cuda for ray-tracing for I believe ground research, so there is definitely already some usefull things happening.

I mean there are a lot of applications for linear algebra, although I admit I don't fully know in what way "AI" uses linear algebra and what other uses overlap with it.

[–] blaue_Fledermaus@mstdn.io 6 points 2 weeks ago (2 children)

They are neat tools, if looked at realistically. They certainly don't deserve to be called AI. I like to call them High Coherence Media Transformers.

[–] Th4tGuyII@fedia.io 6 points 2 weeks ago (1 children)

Considering the "hallucinations" I've seen, they're definitely high something haha

[–] o7___o7@awful.systems 4 points 2 weeks ago

"Decepticons" is right there!

[–] froztbyte@awful.systems 4 points 2 weeks ago (1 children)

we call them autoplag(iarism) machines, much more honest

[–] blaue_Fledermaus@mstdn.io 1 points 2 weeks ago

Yes, that certainly applies to the most popular ones, but not necessarily true for all instances of these technologies.

[–] geneva_convenience@lemmy.ml 15 points 2 weeks ago

But they got an MKBHD ad. Surely MKBHD would not deceive everyone.

[–] Moc@lemmy.world 12 points 2 weeks ago* (last edited 2 weeks ago) (3 children)
[–] homesweethomeMrL@lemmy.world 12 points 2 weeks ago

AI has already taken tens of thousands of jobs. Which it's not doing still.

I hope all the creative teams that got sacked because the idiot CEO blew ten years worth of capital on an AI gamble hold out for C-suite pay to return.

[–] phoneymouse@lemmy.world 3 points 2 weeks ago (2 children)

Feed that picture into ChatGPT and ask it to explain the joke

[–] kn0wmad1c@programming.dev 2 points 2 weeks ago

The one thing LLMs will always be better at doing than humans is pattern detection.

Incidentally, the best job suited to their capability to detect patterns in large amounts of data is doing what CEOs currently do

[–] Xenny@lemmy.world 8 points 2 weeks ago

I thought we were supposed to get like protocol droids with AI. We just get protocols that steal all our data.

[–] Draegur@lemm.ee 8 points 2 weeks ago* (last edited 2 weeks ago) (13 children)

I for one can't wait for Dan Olson (Folding Ideas) to do a feature length deep dive takedown of it soon :D

load more comments (13 replies)
load more comments
view more: next ›