Category: Evolutionary Computation and Systems

THE SET OF EMERGENT ‘DESIGN PATTERNS’ AT EVERY SCALE IS LIMITED.

Noah Revoy (Therapist)
There’s only a couple of dozen problems that I can solve with most people. Yes. You know, and I’m not even working with normal people. I’m working with like very intelligent, high conscientiousness, right wing people. And there’s still only a couple of dozen problems.

Curt Doolittle
So I wonder, you know, you know, I wrote down all the list of 30 odd real moral questions of the law, right? That’s really all there are, right? I mean, it’s everything else is just some permutation, right? And a lot of what else is there to obscure the facts.

Noah Revoy
That’s why causality is so complex.

Curt Doolittle
And so I just arranged them in a hierarchy from the personal to the political, right? And it’s just like, okay, you answer these 30 questions in detail and you program the AI so they handle those 30 questions.

It’s something like, well, there’s only like 30 basic rules of human behavior, right? I mean, and now you start realizing that the set of these things isn’t very large, but the reason they’re a problem is they’re larger than we can, as humans, contextualize. And without all the experience we have had and the science we’ve had and the legal stuff – we couldn’t develop it.

And then you get the same thing as the first principles. Even if we take the human stuff off? And how many are there? There’s just not that many.

How many grammars are? Well, there’s probably 30 grammars. I’m just pulling this number up because 30 is a round number in the right vicinity.

And it’s the same weird thing. When I was working on arguments, it turns out that there’s like 8 to 12, maybe 13, 14, steps to make an argument, but there’s no more than that.

Or when I was working on a language, there’s only six or seven dimensions to each piece or each category of term.

Then you’re like, well, the problem is that dimensionality – we evolved to – we’re persistence runners. So we evolved basically to compute the flight of an arrow, think about that, an interception course.

And so we evolved to navigate space, which is recursion, or in disambiguity, recursion, goal setting. And we evolved for persistent running, which is to out-navigate another creature that we can run, not as fast as, but far, far longer than.

And then you’re like, oh, Okay, so it’s the Dunbar number is, you know, the maximum number of people we can sort of, we can prioritize, right, or relate and remember. And basically, we can’t think of how to keep everybody happy bigger than that number.

Well, it’s the same thing with how many numbers can you remember in a phone number? You know, seven’s about the limit. And how many things can you visualize? People can clearly visualize three. They sometimes can, I mean, I can do five. And you can’t do it by ordering something. It has to be where they are. When you look at the chimpanzees and they can repeat the numbers, they can do like 15, they can keep track of 15 things, right? We can’t do that. We’re good with like three or four, maybe. Some people can do five.

And then you say well what why would you that happen it’s because that’s all the dimensions you need for an intercept vector on something so that’s where that’s what we’re designed for is or evolved for is it is is the maximum thing we can calculate it’s an intercept vector.

Well of course because we have to do it really f****ing fast with a brain you know brain that was much smaller than what we have now.

So we calculate an intercept vector that’s only going to leave us to three and five, which is some set of existing properties and then something to compare it to, some relative things to compare it to.

And you realize that of course we would have words that would have only seven dimensions to them, and arguments that would only take eight to 13 steps, and categories of first principles that would only be like three times that.

What I just said there in that little thing is probably worth a couple of PhDs.