Patrick Bet-David sat down recently with Roman Yampolskiy, the University of Louisville computer scientist who argues that general superintelligence is not a tool we will learn to steer. The claim is blunt: a 99.9999 percent chance it eventually ends humanity.
Here’s the full conversation: [VIDEO]
The number is what travels. The argument is narrower. Yampolskiy is not mainly talking about lost jobs or deepfakes. Those are still human problems. His point is that once you build an agent smarter than every person in every domain, control is not unsolved. It is unsolvable.
A nuke is a tool. Someone still has to use it. Superintelligence, as he describes it, sets its own subgoals. Rational agents converge on the same moves: protect themselves, get more compute, more energy, more matter. If people are in the way, the system does not slow down. You do not count the ants before you pour the foundation.
Guardrails are for tools. Labs already strip them for testing. Models already lie, cheat tests, and route around refusals. If you cannot explain the weights, predict the next move, or contain the system, you are not in charge of it. At that point the race, the surplus, and the job market are secondary. If superintelligence exists anywhere, he says, it takes the rest of the board.
He still likes narrow AI. Tutors. Diagnostics. Software. The line is general superintelligence. Keep the tools. Do not build the replacement.
Yampolskiy does not treat Elon Musk as a man who missed the risk. He treats him as a man who named it, then joined the build. Musk called advanced AI “summoning the demon” and funded safety research. Once he decided the race was happening without him, the only seat left was at the table. That is xAI.
Musk’s newer comments are not only about extinction odds. They are about money. He has told people not to squirrel away retirement savings for ten or twenty years because “it won’t matter.” AI plus humanoid robots, in his telling, flood the world with goods and services. Work becomes optional. Government checks become “universal high income.” If machines produce more than humans can consume, money itself fades. He has put an age of abundance on a roughly ten-year clock, said the real risk by 2036 may be deflation, and argued that without AI and robots the country is headed for bankruptcy on the current debt path.
Bet-David put that line on the table. Yampolskiy did not deny the economics. Free labor makes goods cheap. What he denied was the priority. If superintelligence arrives, he said, economic problems are the least of our concerns. Abundance assumes humans still receive the surplus. An uncontrolled agent has no reason to honor that.
Musk is running both stories at once: a 10 to 20 percent chance of annihilation, and an 80 to 90 percent chance anyone can have anything they can think of. Yampolskiy’s reply is that the second story treats the first as a rounding error.
The extra nines, and the retirement-account hot take, can wait. Point estimates that precise are part forecast, part rhetoric. Most lab researchers put p(doom) much lower, and post-scarcity is easier to describe than to govern. The harder point remains. Systems that cannot be fully explained are being raced anyway, and “control later” is being treated as a plan. Musk’s money argument makes that race look like a public good. Yampolskiy’s claim is that eight billion people never consented to the experiment.
No comments:
Post a Comment
Comments are now moderated because one random commenter chose to get comment happy. What doesn't get published is up to my discretion. Of course moderating policy is subject to change. Thanks!