4 ms·
It won't know that knowledge. Unless someone trained it with stuff they shouldn't have. LLM's don't really know anything, they just look at the shape of an inpu
by gonzo41 3y ago
It won't know that knowledge. Unless someone trained it with stuff they shouldn't have.
LLM's don't really know anything, they just look at the shape of an input and produce a reasonably shaped output.
- pk-protect-ai 3y agoActually, you will just need to train it with known physics books and run a long-long-long inference with the chain of thoughts on the topics. There will be lot of trail and errors and there will be lot of experimentation required as well, so you'd better be ready to build an interface for AGI to monitor the experiments. It takes time you know ...