If AI is going to get out of hand …

A lot of people are predicting that AI will get out of hand and create a nightmare scenario. I don’t know if I believe that or not. Computers are pretty amazing, but I suspect there are limits to what they can do.

However, it’s certainly a possibility we shouldn’t ignore.

Given that, what is there to do about it?

Regulation isn’t going to stop it. First of all, the regulators aren’t smart enough. But even if they were, we can’t regulate every country in the world, nor get every country to agree to regulations.

People have proposed things like …

  1. Ethical guidelines for AI. Who says AI would follow them?
  2. Oversight bodies. That’s like the regulators. They’re not up to the task.
  3. AI kill switches. Who says AI won’t be able to disable the switch?

The solution in Dune was to destroy all “thinking machines” and kill anybody who tried to make one. Up to and including nuking them.

Is there a better option?

Here’s a long but interesting discussion on the future of AI and humanity with Mo Gawdat, who was some bigwig at Google.

Scientific discourse and the magnification of downstream effects

You all probably know that Jordan Peterson has a very strange diet. He only eats beef.

Let’s say I was curious about that, so I asked his doctor “How is Dr. Peterson doing on his all-beef diet?”

The doc could reply with an answer to my question — i.e., he could tell me how this diet is affecting Dr. Peterson.

It’s pretty likely that he would be worried that I would assume that if it’s good for Dr. Peterson, it would be good for me as well, so he might add “but everybody’s different, and just because it works for Dr. Peterson doesn’t mean it will work for you. In fact, it’s very unlikely that this diet is good for most people.”

That seems to be a reasonable level of caution. Experts have to think about how their advice will be used, and that means sometimes they can’t simply give a straight answer to a straight question.

But what if the doctor was also a climate alarmist who believed humanity needed to move to a plant-based diet? He might simply add that.

“an all-beef diet seems to be working for Dr. Peterson, but it’s not a good thing for the planet because [insert ideological baloney].”

That kind of response is skirting the edges of a “reasonable level of caution.” The man’s a doctor, not an expert on climate, or world economies, or any of the things that go into deciding how the world ought to feed itself.

The next step is far worse. Because the doc believes the alarmist conspiracy theories about how methane from cows will destroy the planet, he has to lie about the effect of the diet on Dr. Peterson.

This is where we get into the “noble lie,” where experts tell us things they know are not true because they’re concerned with how the answer will affect downstream issues.

I think this is happening at an increasing rate. Experts are presenting false claims because they’re afraid of the consequences if the truth gets out. They’re also afraid of the consequences to their own careers if they say something that the ideologues believe will cause harmful downstream effects.

I can see two ways to prevent this.

The first is to re-emphasize basic honesty. The second is to provide something like whistleblower protection for dissenters.

Will the Trump indictments ensure he gets the Republican nomination?

At this point in the game, it seems clear that any Republican will be able to beat Biden in 2024 — except Trump. Trump is so toxic that even ultra-lame Joe Biden will be able to beat him. I think a sick dog could beat Joe Biden. But not Trump.

Therefore the Democrats want Trump to be the Republican nominee.

So … are the indictments an attempt to stir up sympathy for Trump to ensure he is the nominee?