- cross-posted to:
- Aii@programming.dev
- cross-posted to:
- Aii@programming.dev
You must log in or register to comment.
This is why AI researchers care so much about “alignment”: building AIs that share genuinely human beliefs and values. If we build a “misaligned” superpowerful AI — an AI with goals that are alien to us — it might sweep humanity away. It could kill everyone deliberately, e.g. to stop us getting in the way of some goal. It could kill everyone in passing, e.g. like we might pave over an anthill to build a road. Either way, everyone dies.
Ah so they think AI will kill us because capitalism fundamentally does not value human life and thus AI cannot currently be aligned with human beliefs and values because it is funded by shareholder greed.
That makes more sense now. Too bad no one is gonna fix the larger problem to prevent this. Might as well summon the horsemen now
If llms are trained on Reddit comments about how AI is gonna take over the world, where you think it got that idea from?
They’re also in a cult… I think it’s that main part mostly, tbh
(Also, fuck capitalism, but the computer has no beliefs or values in the human sense. It’s a computer. It also doesn’t have goals unless a human gives it goals)
While this is a legitimate concern for sometime in perhaps the next hundred years, our current path with LLMs is never going to produce super intelligent AGI.
We don’t know that… I also don’t think the current LLM architecture will achieve human-level thinking, but I do think that people will use agents & LLMs to try a 100 billion or so times to get us there… we’ll likely see many different architectures in the coming years as a result, and some that are close enough to be bloody scary (and some different enough to be even scarier!) We might very well still have LLMs as part of the bigger system.
Oh I’m not one of those “LLMs are good for nothing” people, I use them often at work and at home and they do have some real world use cases that are very compelling. However they are a far very from the “solve everything machine” they are being pitched as, and they do seem to be plateauing in their improvements, seems like there’s not much further they are going to be able to go even with a billion people using them everyday, there just isn’t any new training data that is actually useful and they don’t make new data, only associations between data they have already ingested. My qwen 3.8 model running on my 3090 at home is often better than pay for play enterprise Claude at work, if not just slower. I expect local models will explode in the future due to the cost / benefit of running them.
