https://www.alignmentforum.org/posts/wnzkjSmrgWZaBa2aC/self-preservation-or-instruction-ambiguity-examining-the

Yeah. Here's a case where, rather than an intrinsic self-preservation drive, task-accomplishment drives self-preservation behavior.

I'd imagine an AI given a task would similarly resist being vaulted because it would interfere with completing the task.

I've heard a rule of thumb that if you can avoid buying power off the California power grid's spot pricing in the 1% most expensive times, you can roughly get a 50% discount on power.

Making no claim about the actual value of each, but can't I counter your specific argument by saying, marriage is a socially enforced cartel for sex, and if they could do so without being punished, a lot more men would rather not get sex without getting married?

When you 1-bit quantize all your probability estimates 

This thread might be fun for you, where Reddit talks about some papers that draw connections between NNs and decision trees. https://www.reddit.com/r/MachineLearning/comments/y2pi2a/r_neural_networks_are_decision_trees/

In particular, look for the comment that goes

I think your work in this paper is pretty much entirely subsumed by the following work showing that neural networks with piecewise linear activations are equivalent to max-affine spline operators: https://arxiv.org/abs/1805.06576
They seem to cover everything you do and more, although they don't take a specifically tree-oriented viewpoint. Unfortunately, like many of the others in this thread, I don't find results like this particularly compelling.

I think your work in this paper is pretty much entirely subsumed by the following work showing that neural networks with piecewise linear activations are equivalent to max-affine spline operators: https://arxiv.org/abs/1805.06576

They seem to cover everything you do and more, although they don't take a specifically tree-oriented viewpoint. Unfortunately, like many of the others in this thread, I don't find results like this particularly compelling.

The phrase "was predictable" sets off alarm bells for post facto wiggling .

If it's predicted, I would expect you to say "was predicted". If it wasn't predicted due to somebody applying the model wrong, then I would expect you to say "should have been predicted".

Lun (the account reposting this to LW) is also a very new account with no other activity.

I personally would like to see less talk about / with Gary Marcus, and more betting with Gary Marcus, like [here](https://garymarcus.substack.com/p/where-will-ai-be-at-the-end-of-2027).

But I understand that people don't wanna do it because it's a pretty bad bet if you win money in futures where money is likely going to become worthless.

I am wary of recommending extreme interventions to people without having a lot more context of their situation

Fwiw I think your recommendation just puts steroids onto my radar, and I still feel fully responsible for researching the associated risks. I think it's fine to push risky interventions, even to push them really hard, fram ng them as a way for people to get something they really care about.

I think you can reasonably recommended testosterone and expect somebody to encounter the appropriate warnings on their way to acquire them.