Sam Altman smiles through another “it’s improving” while the model keeps making things up
OpenAI CEO Sam Altman, once fired for being “not consistently candid in his communications,” smiles, acknowledges the problem just enough to sound reasonable, and tries to write off a completely broken and lying LLM as another step on the ladder of progress.
The LLM can not do what it says it is doing. It provides dangerously false information. Altman beams a smile and says this is a “known issue,” and attempts to move past it.
The future is now!
“Real AI” is like full self driving in cars-always just about to arrive. Should it ever get here we are likely to be unhappy.
There's a connection between the law of conservation of energy and, let'say... a law of conservation of information. You can't get additional information by just restating (reshuffling) old information. And that's what is so often running contrary to what the techbros are selling.