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:
I didn’t like the test. I ended up as:
The closest archetype to me sounds more like:
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.
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.