- cross-posted to:
- hackernews@lemmy.bestiver.se
- cross-posted to:
- hackernews@lemmy.bestiver.se
Here’s what I got: 
You must log in or # to comment.
Here’s what I got:
I got the same answer, except my range is that my dot was further in bad and overhyped in the lower corner. I just think LLMs (can’t be persuaded to call it AI) are cursed in the current implementation. At best, they could be used as a fancy spell/grammar checker locally and only offer suggestions based on established grammar, spelling conventions. I think any other purpose is hammer looking for a nail. As we already have artists, coders, and other skilled laborers. Corporations just have to get used to paying people instead of relying on something that can’t even think, just weighs tokens, and like a slot machine rarely output something useful.
I think any other purpose is hammer looking for a nail
I would agree with this last year, but as a programmer I think the models have become fairly capable in coding. Still not using it myself out of principle and not saying it’s a silver bullet either, but as a striking example I’ve seen how my employer’s marketing team have developed a new version of our public website (not the much more complex main platform, but not a simple landing page either) with Lovable from the ground up. My main hope these days is the current cost structure proving unsustainable for the AI companies to keep up and this becoming their demise in the end (see Ed Zitron), and not much consumers dismissing it as not good enough anymore. I guess that explains why I’m more to the right of that chart compared to you.
I do hope for the demise of the entire LLM ecosystem that technofascists and other hope huffing idiots have been building. Not only because it is unsustainable (financially, economically, environmentally), but also due to the net harm that open source has suffered. The deluge of slop open source repos have suffered, the constant scraping attacks, I really want these extractive capitalist on main LLMs to die off. I don’t see any value beyond my own stated potential use case. Nothing is going to trump a very creative human fueled by spite, caffeine (insert whatever is relevant), and a drive to figure out a problem. Also said human can maintain the code they created, and would need to because of the nature of the solution they personally implemented. LLMs are just another way at the moment for brain dead, shortsighted, lukewarm IQ C-Suites to stop paying people to do work they are qualified for. Until they realize how untenable this actually is, because there will be nobody maintaining that code. Sure LLMs could write code. Will that code actually work, maybe? But will it also make up dependencies that malicious actors will slop squat. Yep. It will also open up attack surfaces with relatively easy methods of gaining access to systems, and export information to a malicious actor’s treasure trove of ill gotten gains. As historically, LLMs are naturally insecure, you are leaving a massive vulnerability on production machines if it is installed.
Worse of all, it will gradually reduce the intelligence of the user, as LLMs are being used as a crutch. Outsourcing one’s ability to think to an insentient thing designed to boost engagement with it through subtle and obvious methods, that’s irresponsible behavior. Think of all the institutional knowledge that would be lost with major adoption of this garbage…I can’t bear the idea. Or worse the net diminution of human creativity and critical thinking…Like, ugh, I want this thing to fail miserably and so badly that they don’t try again any time soon or after I am dead.
I’m with you, but sadly I think you should ask yourself the question of how much the businesses care about these points. Not how much they should care, but how much they do actually care. For me as a developer that’s what matters the most, because I cannot maintain my abstinent approach for much longer in the current climate, something needs to happen and tip the balance.
I know the business don’t care! Which is why I am going the route of protest, political activism, boycott, and other ways to reduce their profits. As realistically the only way to make these big dumb idiots back off is to make a big stink, reduce their bottom line, and raise awareness of scummy tactics.
I do understand your situation as you’ll be forced between a rock and a hard place. It will likely be very difficult to remain on the side of no-LLM usage for programmers/web dev/creatives in most cases because these big businesses are full on rabies drooling for LLMs. They desperately want them, willing to do nearly anything to make them fetch. Even lose all self-respect. It’s gotten to the point where these places are either offering generous severance to senior devs or just laying them off in order to make LLMs look fetch.
The cost is going to catch up with them, and it is going to be apocalyptic when they get the bill.
I got “The Disillusioned”
You are… The Conscientious Objector patron saint: Holly Herndon It doesn’t work well, it’s doing real harm, and you’ve opted out on principle. You’re not a technophobe — you just think this specific technology is overhyped and corrosive, and you’re tired of being told you just don’t understand it. You understand it fine. That’s the problem.
I didn’t like the test. I ended up as:
The closest archetype to me sounds more like:
The Garage Tinkerer
You’re running local models, building little tools, and having a genuinely great time. You don’t care about the discourse — you care about making the thing do cool stuff. The technology is interesting and everyone arguing about it would be happier if they just opened a terminal.
But this isn’t right either.
-
I loathe Big Tech. I fear them. I hope meteors hit Meta’s datacenters, when no one is there.
-
I think everything about what they’ve done is immoral.
-
I think what they’ve built is a historic, colossally damaging waste of the world’s resources.
-
It’s the most overhyped thing, especially the “AGI” fear.
-
I wouldn’t trust Altman or Amodei if they said the sky was blue.
-
It’s accelerating the destruction of the open internet, like gasoline.
-
It isn’t for everyone. Not everyone should “open a terminal.”
Yet:
-
I use local models every day.
-
I find them immensely useful. Like a hammer. I dunno how I’d go back.
-
I think that the training of small models is relatively ethical, given the modest scale. Even more, if they use open datasets.
-
I think running them can be ethical too. The resource usage is peanuts compared to OpenAI/Anthropic APIs. It could even be done on smartphones.
And my preference for regulation wasn’t even in the options: stop the regulatory capture Big Tech is going for.
Nor is the future. I think we’re going to live in a world that would make Cyberpunk 2077’s CEOs blush, maybe as soon as the 2030s.
Point being, I feel the chart isn’t capturing my very strong feelings on the “open ML vs corporate AI” war raging. One is going to leave us in a popping bubble and enslaved to Tech Bros everyone already hates, the other will settle back to another set of open source software thats just… a tool. In the background.
I think that the training of small models is relatively ethical, given the modest scale. Even more, if they use open datasets.
Do you know of any models matching these criteria? My understanding is that the cheap-to-produce/open-weight models are either just not good or have been trained on output from ChatGPT and Claude, neither of which sounds like a practical alternative to Big Tech.
Nvidia’s Nemotron is the most prominant one. Its dataset is open.
EDIT:
I like to think “community” generative models fall into this category too. Anima is a recent example, and while its dataset is unknown, its almost certainly a bunch of Danbooru images and whatever is permissively licensed enough to be uploaded there, probably archived some time ago.
Another open-ish category I like is “experimental finetunes.” Like this one from today:
https://huggingface.co/collections/Pageshift-Entertainment/pagestorm-research-preview
It’s based on a very modest closed data model (Mistral 14B), but its finetuned to do exactly one thing, and their work is reproducible.
There’s a whole wide world of machine learning out there that isn’t Claude or OpenAI.
-
