aSanchezStern

I agree that "proof of thought" is a misleading name, but this whole "computers can't think" thing is making LLM skepticism seem very unscientific. There is no universally agreed upon objective definition of what it means to be able to "think" or how you would measure such a thing. The definition that these types of positions seem to rely upon is "a thing that only humans can do", which is obviously a circular one that isn't useful.

Sure, but when you're walking all the time, none of that time is wasted, because you're helping your body and brain function better. When you use a car, you really are wasting all your transportation time. To get the same benefits, you would have to drive places, and then go walking recreationally after, which would clearly take much more time to get the same utility.

Cite?

I don't think that "arguing that something is against the rules" is in the CIA sabotage manual, because it's not generally considered sabotage. Maybe if you argue things are against the rules that you know aren't, to slow things down?

The answer it seems is, it depends on what kind of code you're looking at. The post showed that `for` loops cause a lot more variable-name-biased reasoning, while `ifs` and function defs/calls are more variable-name independent.

Yeah that's linked a couple of times in the post

Thanks for pointing that out, it's fixed now.

If you're interested in a more scientific treatment of the topic, the post links to a technical report which reports the numbers in detail. This post is instead an attempt to explain the topics to a more general audience, so digging into the weeds isn't very useful.

This post actually mostly uses the subset of Python where nullability is checked. The point is not to introduce new LLM capabilities, but to understand more about how existing LLMs are reasoning about code.

Many people don't think we have any good evidence that our brains aren't essentially the same thing: a stochastic statistical model that produces outputs based on inputs.

Yeah the link title is overclaiming a bit, the actual post title doesn't make such a general claim, and the post itself examines several specific models and compares their understanding.

The post includes this caveat. Depending on your philosophical position about sentience you might say that LLMs can't possibly "understand" anything, and the post isn't trying to have that argument. But to the extent that an LLM can "understand" anything, you can study its understanding of nullability.

Where would the "paper" part come in? Is that just based on the word "bench" in general?

"Didn't even win the democratic primary" might be a bit of a misnomer if it's implying that winning the democratic primary should be much easier than winning the general. Bernie has positions that are popular outside of democrats, many people think that if he had won the primary, he would have had a much better chance of winning than Hilary or Biden, since he had less baggage and his policies were broadly popular.

I mean, yeah, they are saying that you can commit a crime and they won't punish you. That's what "suspending enforcement" means. I think what you mean is that "law and order" is usually applied to poor criminals, where as what is happening here is refusing to enforce laws on the rich criminals.