LessWrong 2.0 Reader

Some conclusions should be drawn from  existing countries which use the death penalty well, example Singapore. Low crime is great!

I think the billion-dollar question is, what is the relationship between these two perspectives? For example, a simplistic approach would be to see cognitive visualization as some sort of Monte Carlo version of spreadsheet epistemology. I think that's wrong, but the correct alternative is less clear. Maybe something involving LDSL, but LDSL seems far from the whole story.

First, it's uncomputable to measure performance because that involves the Solomonoff prior. You can approximate it if you know some bits of Chaitin's constant, but that brings a penalty into the description complexity.

Second, I think that saying that comparison is computable means that the utility is only allowed to depend on a finite number of time steps, it rules out even geometric time discount. For such utility functions, the optimal policy has finite description complexity, so g is upper bounded. I doubt that's useful.

I roughly agree, but it seems very robustly established in practice that the training-validation distinction is better than just having a training objective, even though your argument mostly applies just as well to the standard ML setup.

You point out an important difference which is that our ‘validation metrics’ might be quite weak compared to most cases, but I still think it’s clearly much better to use some things for validation than training.

Like, I think there are things that are easy to train away but hard/slow to validate away (just like when training an image classifier you could in principle memorise the validation set, but it would take a ridiculous amount of hyperparameter optimisation).

One example might be if we have interp methods that measure correlates of scheming. Incredibly easy to train away, still possible to validate away but probably harder enough that ratio of non-schemers you get is higher than if trained against it, which wouldn’t affect the ratio at all.

A separate argument is that I’m think if you just do random search over training ideas, rejecting if they don’t get a certain validation score, you actually don’t goodhart at all. Might put that argument in a top level post.

We could have the exact same court procedures, just change the law on the books from 3–5 years to 3–5 fingers.

We could, but with actual humans, we won't.

Or are you just stating, without any justification

"By observing human beings" is not "without any justification". We know what societies that mutilate prisoners are like, because plenty of them have existed.

Also, individuals don't have to "become irrational" for the ones who are already irrational to gain more influence.

You won't find systems with none at all, but you can easily find systems with those aspects to a significantly larger or smaller degree.

But what you're probably not aware of is that 0.8% of the US population ends up dieing due to intentional homicide

That is an insane statistic. According to a bit of googling this indeed seems plausible, but would still be interested in your source if you can provide it.

It looks like the comedian whose clip you linked has a podcast:

https://www.joshjohnsoncomedy.com/podcasts

I don't see any guests in their podcast history, but maybe someone could invite him on a different podcast? His website lists appearances on other podcasts. I figure it's worth trying stuff like this for VoI.

I think people should emphasize more the rate of improvement in this technology. Analogous to early days of COVID -- it's not where we are that's worrisome; it's where we're headed.

We would obviously have to significantly streamline the process, such that people are executed within 6 months of being caught or so.

This is one of the biggest hurdles, IMO. How do you significantly streamline the process without destroying due process? In the US, this would require a complete overhaul of the criminal justice system to be feasible.

Your OP is completely misleading if you're using plain GWAS!

GWAS is an association -- that's what the A stands for. Association is not causation. Anything that correlates with IQ (eg melanin) can show up in a GWAS for IQ. You're gonna end up editing embryos to have lower melanin and claiming their IQ is 150