I've found looking for available products to be one of ChatGPT's most finicky use cases, it consistently recommends products that are unavailable, out of stock, or have wildly different reported prices than what it claims.
> If anything modern humans' most emergent quality is their inability to deal with any serious problem.
Yes. If this problem had come up in, say, 1956, it would have been dealt with.
The generation that fought in WWII had a low tolerance for bullshit, and corrupt or incompetent leaders.
The nearer term worry is not AI wiping out humanity. It's AI wiping out jobs in bulk.
That looks likely at this point.
After that, it's AGI empowering individuals or small groups to cause trouble on a scale that previously required a nation state. AI assisted hacking is working all too well. Fortunately AI can help on the defensive side, and holes are being plugged in bulk. There's a risk on the bio side, but people in the bio field seem not to be too worried about it.
Humanity has been very lucky in some ways. One of the Manhattan Project physicists, discussing the unreasonably difficult problem of isotope separation, remarked that maybe someone will figure out how to do this in the kitchen sink. Nobody ever did, or if they did, it was buried very well. The amount of specially built industrial plant needed to separate uranium is huge.
That's the moat for nuclear proliferation.
That's the only reason every country with an army doesn't have nuclear weapons.
So you’re saying that the actual problem is us debating how AGI will destroy us (because doing so leads to AGI destroying us). How do you propose to deal with that problem?
reply