Jun 28 JDN 2461220

What do you suppose are the chances that, within the next 10 years, AGI is invented and ushers in a new era of abundance in which hardly anyone has to work and we all live in prosperity?

A lot of people seem to think that this is actually pretty likely, for some reason. They seem to have bought into some ridiculous hype by AI corporations. Yes, computing power grows exponentially; but just having enough computing power isn’t enough to actually create artificial general intelligence. You have to actually understand cognition, and we absolutely do not, not even a little bit. Basically at this point we’re just hoping we can throw enough computing power at the problem that we don’t need to understand anything. (And, to be fair, that has worked much better for LLMs than I would have expected.)

But okay, it’s not impossible. I am certainly not going to call it my retirement strategy, but I will allow that there is some chance that this could happen. Let’s go with… 0.1%. I am 99.9% sure it will not happen.

This got me thinking about other extremely remote possibilities, like winning the lottery jackpot. How do those two events compare? Would it be more or less crazy to bank my retirement on the Singularity than to bank my retirement on winning the lottery?

Amazingly, it would in fact be less crazy. By quite a lot.

That’s how bad your odds of winning the lottery are.

The odds of winning either the Mega Millions or the Powerball jackpots (don’t ask me why we have two) are about 1 in 300 million. This works out so that if literally everyone in America played, someone would probably win each time; and it also works out so that with the large number of people who do play, somebody wins every month or two.

But as far as your individual odds of winning, it means that, for all practical intents and purposes, you will not win.

I have long tried to get this into people’s heads, but I think the Singularity comparison helps a great deal. I have said that I think there is a 0.1% chance the Singularity will happen in the next 10 years. That is pretty much not going to happen, right? I’m just being generous and saying it’s possible. A probability of 0.1% is still pretty much negligible.

And yet, to have a 0.1% chance of winning the lottery, you would need to play 300,000 times.

At $2 per ticket, this would cost you $600,000. To do it within 10 years, you would need to spend $60,000 per year on nothing but lottery tickets.

And that’s to get a chance that we just agreed is basically still negligible.

If you spend a “reasonable” amount on lottery tickets, say buying two tickets each week, you only spend $200 a year on it, which for most people is not really a big deal. But your odds of winning within 10 years in this scenario are now 1 in 300,000—that’s 300 times smaller than the odds of not needing to win because the Singularity has happened by then.

Don’t play the lottery. Save that $200 and put it in stocks. $200 a year won’t make you rich, but neither will the lottery—and at least you’ll have something to show for it later.