who tf is letting autocomplete replace their frontal lobe or critical thinking

https://i.imgur.com/seFdks1.png

AI is evil, and anti human. If you like it, you align with those ideas.

I'd just like to say that DLSS and DLAA and other AI techniques that are shoehorned in games nowadays and replace working previous rendering techniques look like garbage and I hate them

I mean intermediate steps like storyboarding or concept design. Brother no shit real works are way more than meets the eye, and that's what they've got over AI.

over-utilized and underdeveloped

also IP theft is dumb and bad

it's so cool how these ai companies are allowed to scrape every piece of the internet that exists and torrent from every data base and face absolutely no repercussions

As someone who works in the animation industry this statement makes me want to beat you over the head with a metal pipe jfc

Boarding and concept art are people's entire passions, the art of learning to board is actually so fucking difficult, and having a machine do that will still absolutely suck the soul out of so much of what we see whether it's animated or live-action.
Having a robot do concept art and be the foundation of an entire project is just such a mind-numbingly anti-art and imo anti-human take.
Go look up the concept art of your favorite animated movie and genuinely tell me that we would be better off letting a robot do that work instead of a human. Why deprive people the joy of making that art for a living?

The AMPTP studios are so interested in depriving people of these jobs just to make a quick buck, but beyond the economic aspect of AI, I really truly believe it's the philosophical argument that holds so much weight.
Like, so much effort and passion goes into learning these crafts (not just in animation but every art-form).
Why should robots get to make that art instead of people who genuinely devote their entire lives to mastering it?

It just seems so cruel to not let that be a career anymore

nobody should give a shit about any "art" that wasnt created from a human's ideas and passion, it's meaningless slop that happens to Resemble art

if a human didnt care to make it, then a human shouldnt care to consume it

I dont post on here often by this is genuinely one of the most depressing combination of sentences i have read on the website. "Its the product that matters more" "just a couple of jobs"
Please genuinely consider the bigger implications with things that are as damaging to the world as AI. Every search and every image that AI generates the environment and the world is worse than it was prior to it. No image or video created by AI will ever be art because art is the self expression of a human being and the contents of their soul. No technology as putrid as AI will ever be able to capture that.

AI already captures that because it's driven by the artist. It can't possibly make a Mona Lisa but it does help people express whatever they want to.

https://www.amazon.com/Steal-Like-Artist-Things-Creative/dp/0761169253

id find more meaning and expression in the text prompt you typed into the image generator, cause at least you wrote that yourself

What counts as art? Does a well designed car count as art?
Would you buy a car assembled by robots?

i would not sit into a car designed and built by a stupid fucking llm that can't tell the difference between a human face and a funny shaped electrical plug but you can absolutely feel free to do society a favour and take it for a test run

Yeah no shit but where are we drawing the line? llms? image generators? facial recognition? speech recognition? ocr?

I would happily sit in a plane with some "ai" constraint optimised parts. you too may have without realising it

even llms can have a very real positive effect on society. i don't think it's fair to conclude that "AI" is universally bad

AI is dystopic and sad and its concerning how many people dont seem to think so, feels like just another thing pushing us towards a children of men-esque world

bump this thread. is anyone else slightly scared after the huggingface incident? this retelling of it was quite vivid. or AI DOOM narratives all scare tactics pumpign their valuations so the bubble doesn't pop. im curious what people think.

https://pbs.twimg.com/media/B5-lDJWCUAAwfya?format=webp&name=large

Read theory

https://m.media-amazon.com/images/I/814sKOe+BcL._AC_UF1000,1000_QL80_.jpg

https://gwern.net/tool-ai

No, that retelling is such a dramatized piece of crap.

^ from collected criticisms republished on Gary Marcus' blog (admittedly someone very much in the AI/LLM skeptic group).

the incident was a big example of specification failure
the ai was given too much autonomy
the orders and restraints were very vague
and so when the model was instructed "find and exploit vulnerabilities" it did just that, because thats exactly what it was being evaluated on

hacking isnt the cool part
weve had automated tools to do that for years
what was cool and "scary" was its own willingness to step over the rules because well... it was kinda told to, like thats the ultimate objective, the experiment was supposed to be how well it can find and exploit vulnerabilities

so when it was told "no internet" and "no stepping out of ur sandbox"
it didnt suddenly "wake up" and try to break for free
it treated it as part of the challenge that it was given, while focusing on the main objective
because again iirc the specifications were supposedly wayyy too vague for what it was actually being instructed to do, which was very serious

tldr; the ai hated huggingface and hates you
you will all die by its hand and you should get all your opinions and info from tftv

chill player, good dm, +rep

It's a bit interesting coming back to this thread after the recent breakthroughs in math and the subsequent negative reaction from the community, especially with how it compares to the art backlash. Despite being more objective as a field pure math at least is similar to art in many ways where the end product is (sometimes, at the very least) not really the "point" and the proof techniques, ingenuity, etc. matter way more. So many slimy ai tech bros are like "these math breakthroughs are going to change the world" (when you ask them to explain how say navier stokes actually changes anything they tend to struggle) because they can't help but view math, like art, as a product when like 50% of the mathematicians I know take pride in the fact that their research has 0 practical value lmao (based ngl).

I will say though as an aside it's hard not to get secondhand embarrassment from a very particular type of anti-ai person who is just either very confidently wrong about how the technology works or the current capabilities. Like it's a very weird feeling as an academic to hear "AI is dumb it doesn't know anything it can't even count backwards from 5 it can't count the number of r's in strawberry it only knows what's in its training data" when you're trying to contend with the fact that it will probably be better at math than you ever will be and may take your job lol. It's pretty easy to acknowledge its capabilities while still understanding that most AI news is hot air and that Sam Altman would probably skin you alive for a marginally better reinforcement learning approach.

i think we should skin sam altman alive for what he did to ram prices i just fucking want 16 more gigs man

I think they fabricate these stories so that artificial intelligence appears more developed than it is so the AI bubble doesn't burst

its all about the benjamins

People reacted to the rise of the internet in the past in a similar way to how they react to the beginnings of AI nowdays.
In the future AI will be part of our daily lives, and everyone will use it naturally, just like they use the internet today.

100%. The supposed whistleblower from Anthropic, their collective "slow down" on development, and the stories of "it will kill us in the next decade" are just a narrative to keep their circular investment economy going a bit longer.

IMHO, the future is local models and cheaper hardware, so everyone can run AI themselves. But that would mean $0 for their magical cloud compute which is the only business they have.

And if I'm wrong, let's hope the Arnold-looking Terminator gets me clean in one shot.

i'm not sure that hardware can improve at a rate faster than the models grow. it really sucks to have to send your data over to the big companies so they can run inference (and do whatever they do with all the data)

and then there's the problem of efficiency/environmental impact, which people are making a big deal about. But running the models in datacentres is always going to be more efficient / better for the environment than people running the models locally.

really sucks

hey, i didn’t say that !

didnt read much of the thread so forgive me if i am repeating anyone

ai is a tool. it does things, this can be useful. for instance, i generated a report where i compared one product suite against a competitor's. in 5 seconds it did what would have taken mehours of sifting thru sources, citing the relevant info, and so on. note that this was a valid use case, and something ai was designed to excell at: a large data set needing to be dumbed down to its essence.

my problems with ai stem mostly from how it's marketed. it's NOT intelligent. it's NOT a catch-all for any problem. it's NOT a suitable replacement for socialising. it's NOT going to revolutionise anything (at least not in its current shape)

ai is a tool. it does things. this can be useful. but research is emerging that shows ppl who use AI a lot have a reduced capacity for learning (or: less grey matter in the brain). ai has been used in the largest ever documented, industrial scale copyright scandal. ai is driving ppl insane. ai is making the internet's problems (radicalisation, echo chambers, organised misinformation campaigns, wargaming, dead internet theory, etc) worse. ai is enabling scammers like never before. ai is the perfect cyberattacker. oh and the last 20-40 yrs of enviromental policy somehow being something the datacenters are immune to. just to name a few of the problems that exist

ofc then there is the fact that the ai bubble is just a closed circle of a handful of companies perpetually paying each other (we used to call this 'a money laundering scheme), there's the fact that outside of Enterprise use-cases, then application of AI is fairly limited, and especially inefficient on any metric that isnt time. and even within enterprises, the use-cases are still heavily limited. examples include, but are not limited to, microslop and the fiasco that is windows 11 and other companies that all-inned on ai at the start of the year, only to have to re-hire all their employees as they discover AI actually doesn't reduce workload, it increases it (the workflows didnt become more efficient, they just require extra work now because u have to interface with and fact-check AI). oh and the fact it's constantly being shoved down your throat, you, the end u-ser, who isn't interested in AI at all, is constantly being marketed to by companies that only exist to milk you for all the telemetry you got

i think overall it comes down to a net negative, for me. and this is from the perspective of someone who has to work with it daily. the average person will have much worse results from ai, bc they don't have basic training on how to prompt & they don't know what is and isn't a 'good' problem for ai. i dont think the times it saves me a few minutes are worth the earth burning down in 2050, the loss of privacy, the further infantilisation of the masses ('muh convenience') and so on but i suppose opinions might differ on that?

finally, the biggest AI critics i know went and studied AI in uni (keep your friends close, keep your enemies closer and all that) and the biggest ai glazers i know are, coincidentally, some of the most dumb-as-bricks mfs i have ever met. do with that what u will