[25:41] well, my favorite book is The Fabric of Reality by David Deutsch. So, I think that still holds. I'd hope to answer the questions in that book with with the AGI. That's my post AGI work.Hassabis is a genius, and certainly one of the world's leading AI researchers. But I am lowering my opinion of him, if this is his favorite book.
At first glance, the book is a celebration of modern science. It is enthusiastic about science, theory, the scientific method, explanations, and testability. It explains what science is all about.
Except that it is really anti-science. For all his talk about the merits to testing theories, he mostly promotes theories that cannot be tested at all. For all his talk about reality, he mainly believes in things that are not real.
His favorite ideas are Popper's falsification, Darwin's evolution, and many-worlds quantum mechanics. He raves about their explanatory power.
These are so bizarre, because many-worlds is just an unscientific belief that cannot be falsified.
Darwin evolution has some explanatory power, but the main principle -- survival of the fittest -- is a tautology. Fitness is defined as what survives.
Deutsch starts the book by attacking Steven Weinberg for saying an instrumentalist remark: "The important thing is to be able to make predictions". Weinberg's point was anti-geometry, not instrumentalism, and he later changed his mind when geometry proved crucial for theoretical physics.
For a more up-to-date description of Deutsch's opinions, see this new interview of him. He says that he now believes in free will, because we have the ability to create novelty.
Considering how much of a systematic grift of other people's data AI already is, and how much it bullshits/hallucinates with a straight face, I'm not sure I see how this promotes free will and the ability to create novelty when the AI can steal it faster than you can register.
ReplyDeleteThe last thing I consider something that replaces human creativity and people thinking for themselves is genius.