80
submitted 1 month ago by [email protected] to c/[email protected]
all 30 comments
sorted by: hot top new old
[-] [email protected] 57 points 1 month ago

"We had been working with their AI tool for a while, and it was absolutely not at the point of being capable of writing lessons without humans.”

Lol, LMAO this is gonna be a disaster once the bubble crashes and current models undergo recursive degradation.

[-] [email protected] 39 points 1 month ago

I do think LLMs are going to start getting worse once more data is fed into them. And then instead of admitting this. These companies will have so much capital they will just tune the AI to say exactly what they want it to every time. We already see some of this but it will get worse.

[-] [email protected] 15 points 1 month ago

Reject LLMs branded as AI, retvrn to 9999999 nested if statements.

[-] [email protected] 8 points 1 month ago

LLMs are going to start getting worse once more data is fed into them

I've been hearing this for a while, has it started happening yet?

And if it did, is there any reason why people couldn't switch back to an older version?

[-] [email protected] 17 points 1 month ago* (last edited 1 month ago)

I knew the answer was "Yes" but it took me a fuckin while to find the actual sources again

https://arxiv.org/pdf/2307.01850 https://www.nature.com/articles/s41586-024-07566-y

the term is "Model collapse" or "model autophagy disorder" and any generative model is susceptible to it

as to why it has not happened too much yet: Curated datasets of human generated content with minimal AI content If it does: You could switch to an older version, yes, but to train new models with any new information past a certain point you would need to update the dataset while (ideally) introducing as little AI content as possible, which I think is becoming intractable with the widespread deployment of generative models.

[-] [email protected] 8 points 1 month ago

The witting or unwitting use of synthetic data to train generative models departs from standard AI training practice in one important respect: repeating this process for generation after generation of models forms an autophagous (“self-consuming”) loop. As Figure 3 details, different autophagous loop variations arise depending on how existing real and synthetic data are combined into future training sets. Additional variations arise depending on how the synthetic data is generated. For instance, practitioners or algorithms will often introduce a sampling bias by manually “cherry picking” synthesized data to trade off perceptual quality (i.e., the images/texts “look/sound good”) vs. diversity (i.e., many different “types” of images/texts are generated). The informal concepts of quality and diversity are closely related to the statistical metrics of precision and recall, respectively [39 ]. If synthetic data, biased or not, is already in our training datasets today, then autophagous loops are all but inevitable in the future.

[-] [email protected] 7 points 1 month ago

Sometimes yeah you can see it. Not only with updates but within a conversation, Models degrade in effectiveness long before context window is reached. Things like image generation tend to get worse after >2 edits and even if the image seed is given .

[-] [email protected] 44 points 1 month ago

When the AI bubble pops, it will take the US with it.

Good.

[-] [email protected] 44 points 1 month ago

Ah yes, the bourgeoisie really know their game if they let young adults unemployed in masse for a while. Not at all linked to social transformations.

porky-happy

[-] [email protected] 43 points 1 month ago

This AI shit is going to deskill us back to the stone age LOL. Real subsumption of thought.

[-] [email protected] 19 points 1 month ago

subsumption of thought into models that dont think agony-limitless

[-] [email protected] 13 points 1 month ago* (last edited 1 month ago)

the ceo of the company I work at was just complaining to me and my colleague earlier today that his software engineers aren't using enough AI to do their work....

[-] [email protected] 42 points 1 month ago

Can we stop talking about Duolingo? I know why all these online writers talk about, it because tech employees thought they found one of those cushy jobs for life and now the rug is being pulled out and it's example A. However, automation taking over jobs at ports is a much more large scale issue and is an example of how automation will infiltrate every sector not only programmers.

[-] [email protected] 39 points 1 month ago

Unfortunately I don’t expect Americans to realize that C suite jobs can just as easily be replaced by AI. We better keep on anthropomorphizing this shit and not treating it like a tool!

[-] [email protected] 35 points 1 month ago* (last edited 1 month ago)

C-suite jobs are the jobs most easily replaced by AI, and arguably the only class of job that current AI models are capable of performing adequately.

They're task-minimal jobs that don't require any real understanding of anything. Basic logical decision-making and long-term analysis can outperform the typically self-centered and petty motivations most C-suites apply to make decisions that maximize present-quarter profits over everything else.

[-] [email protected] 33 points 1 month ago

Jira can already be programmed to annoy me about open issues that have not had an update in a couple days. Program it to message me at inconvenient times asking me stupid questions about the project that I've been working on for 2 years, program it to deny me a raise, and you could replace half of the middle managers at my company.

[-] [email protected] 32 points 1 month ago

Learn to ~~code!~~ farm!

[-] [email protected] 31 points 1 month ago

I can't wait for the next AI winter.

[-] [email protected] 29 points 1 month ago

The US climate getting increasingly hotter with younger, alienated men becoming increasingly jobless. Looks like you’re speed running for the Guatemalan experience. Hope you enjoy the increase in violent crime.

[-] [email protected] 2 points 1 month ago

Trumps telling them that it’s the illegal immigrant's fault, so that should make it better.

[-] [email protected] 28 points 1 month ago

I'm skeptical given the downward trend on that graph started well before winter 2022 (when ChatGPT first came out), and even before covid.

[-] [email protected] 20 points 1 month ago

We never recovered from the crash of 2008.

[-] [email protected] 28 points 1 month ago

Just reeks of desperation from the bourgeoisie

[-] [email protected] 25 points 1 month ago
[-] [email protected] 20 points 1 month ago

I don't like it.

[-] [email protected] 5 points 1 month ago

Chad Anki > Virgin Duolingo

[-] [email protected] 15 points 1 month ago

But I just read an article from techbro.mag.richbootlickers.org that said I was only imagining this and AI will make a perfect utopia for everyone forever!

[-] [email protected] 14 points 1 month ago

Mr Merchant, please lead the luddite uprising against the chat bots

[-] [email protected] 10 points 1 month ago* (last edited 1 month ago)

The Butlerian Jihad, also known as the Great Revolt as well as commonly shortened to the Jihad, was the crusade against computers, thinking machines, and conscious robots.

this post was submitted on 02 May 2025
80 points (98.8% liked)

news

24084 readers
653 users here now

Welcome to c/news! Please read the Hexbear Code of Conduct and remember... we're all comrades here.

Rules:

-- PLEASE KEEP POST TITLES INFORMATIVE --

-- Overly editorialized titles, particularly if they link to opinion pieces, may get your post removed. --

-- All posts must include a link to their source. Screenshots are fine IF you include the link in the post body. --

-- If you are citing a twitter post as news please include not just the twitter.com in your links but also nitter.net (or another Nitter instance). There is also a Firefox extension that can redirect Twitter links to a Nitter instance: https://addons.mozilla.org/en-US/firefox/addon/libredirect/ or archive them as you would any other reactionary source using e.g. https://archive.today/ . Twitter screenshots still need to be sourced or they will be removed --

-- Mass tagging comm moderators across multiple posts like a broken markov chain bot will result in a comm ban--

-- Repeated consecutive posting of reactionary sources, fake news, misleading / outdated news, false alarms over ghoul deaths, and/or shitposts will result in a comm ban.--

-- Neglecting to use content warnings or NSFW when dealing with disturbing content will be removed until in compliance. Users who are consecutively reported due to failing to use content warnings or NSFW tags when commenting on or posting disturbing content will result in the user being banned. --

-- Using April 1st as an excuse to post fake headlines, like the resurrection of Kissinger while he is still fortunately dead, will result in the poster being thrown in the gamer gulag and be sentenced to play and beat trashy mobile games like 'Raid: Shadow Legends' in order to be rehabilitated back into general society. --

founded 4 years ago
MODERATORS