@aihorde@lemmy.dbzer0.com draw for me Lemmy user ComradeMiao prompting an ai to create an image of the db0 admin as a power hungry reddit mod
This could be a good use of AI. Since this regime is doing it, and since some of their claims are pretty unrealistic, it probably won't be. But, ML has been used for a while to help identify new drug compounds, find interactions, etc. It could be very useful in the FDA's work - I'm honestly surprised to hear that they're only just now considering using it.
The Four Thieves Vinegar Collective uses some software from MIT ASKCOS that uses neural networks to help identify reactions and retrosynthesis chains to produce chemical compounds using cheap, homemade bioreactors. Famously, they are doing this to make mifepristone available for people in areas of the US without access to abortion care.
You can check it out here. It's a good example of a very positive use-case for an AI/ML tool in medicine.
Only if it's wheelchair accessible though
Also are ignorant of minor (and major) details
Disregarding the /s bc i want to rant
I guess if you described board states in language and got them to recognize chess board states from images (by describing them in language), and trained them on real games, you could probably make a really inefficient chess bot.
But that said, you could use an "agentic" model with an mcp to route queries about chess to an api that links the LLM to an actual chess bot.
Then it'd just be like going to the chess bot website and entering the board states to get the next move. No magic involved, just automated interaction with an api. The hype and fear and mysticism around llms bugs me. The concepts behind how they work aren't hard, just convoluted
They restored the names, but iirc they named them after random other ppl with the same names. Like fort Hood in TX - now named after Robert B Hood, a WWI colonel. So, he couldn't even do that right and all the chuds in favor of this didn't even really get what they want
I mean they aren't large chess models. They can only do language tasks. They don't think, they predict words based on context and its similarity to the corpus they're trained on.
Yeah, the courts being captured is possibly the biggest problem facing the US right now. They're what makes all the bad stuff the current regime is doing possible.
+1. It was really good
That's a fair point. But it would need to be argued in the courts.
Irony & Stephen being an arch-ghoul notwithstanding, federal law does supercede state law (Art 6 Sec 2). Like, Nebraska couldn't just legalize murder. They could stop enforcing that law, but it'd still be illegal and the federal gov't could take action.
There's a cloudflare outage happening now. It's affecting basically everyone at my job. Idk how widespread it is