cgearhart

Thanks! This is exactly the kind of analysis I was hoping to hear.

Is there anything interesting to learn from the agent written bots? Or are they just inferior versions of the best bots because the time limit is so short? Is there any indication that they’re trained on the code from the human written bots (like, do they magically “guess” creating functionality that requires experimental feedback or domain experience?) How much are the bots on rails applying the same techniques every time?

I think those kinds of questions would be more interesting than the raw win/loss ratios.

I still remain hopeful that this is a step change in capability and capacity that we can adopt and adapt towards our own purposes. What was taking decades for us to do has been reduced to days and dollars. So while we’re witnessing a forest fire consuming entire fields of open problems in math, the best upside would be finding new problems that come from learning how to work with the new tools.

This is what it seems like so far for software. The speed limit has changed. It’s interesting to explore what that means. Extending the analogy, long ago you could be a race car driver part time; that would be pretty laughable for an F1 driver. It’s a very specialized skill and the optimal athlete is quite different for driving a car than for running, for example. That feels like what I’m seeing in software.

The most effective folks are much faster and better with these tools than average. It’s a different skill set.

I think that formal verification makes Math much closer to code than many other fields. But I’m eager to see the first mathematicians that are dramatically more proficient with the new tools than average.

This has been going on for awhile. There was a fella that got flagged by facial recognition in a casino. He was hauled back by security and arrested by the police for trespassing. Even though he gave police his license. They thought it was more likely a fake id than than the “99% match” from the software was wrong. And no accountability afterwards; no change in policy or training.

I hadn’t seen this before, but he’s near the top of my list as an example of this. One the common characteristics I see of people with this mindset is that they have high risk tolerance. Things are replaceable, and they feel safe exploring that trade space.

That’s mostly where I’ve wound up as an adult. I think a lot of it comes down to financial stability. I spent a lot of my own money on books when I was younger. They embodied that cost in perceived value. But I learned that they might have more value than their cost if you’re willing to make them more useful.

We studied the book in order starting from chapter 1 in the first calc class up to the last chapters in calc 4. I was in the last class when he gave me the idea to cut it, so I cut out the last few chapters and carried a very thin volume of what I needed for the last class.

20ish years ago I complained to my college roommate that my math book was too big. He asked if I was gonna sell it back, and I said I couldn’t—they had a new edition out, so mine was worthless. Cut it, he said.

It was something I had never considered. Not for a moment. Books were sacred to me. Cut one? Blasphemy.

But he was right. And so I cut it. And it fixed all of my complaints.

What I learned is that some folks have a different mindset; they see the world as something to be shaped around themselves, rather than seeing themselves as something that should be shaped to fit the world. That’s been a useful lesson sometimes.

Back in the mid 90’s I got it to “run” on a 386, but it was miserable. It was basically wondering if it was hung or just really slow on every command or mouse click.

Out of curiosity, do you have other ideas or things you want it to do? It just seems like they always converge on the same set of executive tasks because they’re a nail that is exceptionally well suited to this particular hammer.

At least in the write up he talked about trying with 3 different models. He could look at the progress at each step and evaluate whether it was making progress or stuck. He was able to understand the partial and complete solutions.

I think part of what it shows is how much a good value function provides in a search problem. But we’ve always known that brute force is also a valid solution to search problems. It seems that what’s changed is that it’s much easier than before to run brute force, so it’s going to be applied more often and by a base rate argument “solve” more problems. What remains to be seen is whether human + new tech = better than the parts, as has been the case with most previous new technologies.

I understand why “prompt kiddie” feels accurate, but I don’t think it is. Expertise is _amplified_ with LLM agents. The same $300 of tokens given to my plumber—who is an _excellent_ plumber—is unlikely to produce the same outcome.

This is the exact opposite of what I’ve been dealing with for awhile. LLMs absolute cannot work on something without an understanding unless they can outsource the understanding to a verifier. If you’ve got an easy to check function to measure progress then “keep going” is all the prompt you need. But if you need it to figure out “I pushed the up button and it moved up and left” then it’ll find the same bug five ways without realizing it’s just one bug in the underlying math.

Zero dependencies should not be the goal. It’s just a modern retelling of “not invented here” syndrome. It’s an anti-pattern because now you own maintenance on something that is not your core problem. In every case I’ve seen so far, the problem is that the person running the LLM doesn’t understand what problems the library solves and so they erroneously think the dependency and their generated solution are equivalent.

This is exactly what I’m talking about. I’m seeing the same pattern. And the most useful proxy I’ve found so far for “we don’t understand the problem” is taking zero dependencies. It means that you haven’t bothered to understand what the library does and you think it can be replaced with something AI one-shots.