My manager doesn't stop overworking. When told on peer performance review that we have people who are consistently overwork because they are swamped, he played it down.

But hey, at least he doesn't encourage overworking either.

  Why are people still upvoting obvious AI slop garbage?

  It’s literally impossible to catch Pac-Man?

  You do not move fast enough

  the Pac-Man AI is programmed for perfection

I’m not going to comment on the code (I abandoned the page early due to the terrible controls) but there’s no contradiction in that part you’re commenting on. “Slop” doesn’t just mean “doesn’t work”; “programmed for perfection so it does not make deliberate mistakes” doesn’t mean “the code is perfect”.

To output text in a terminal you could use `echo`, but you could also write a 500 line function which does a bunch of unrelated garbage then finally outputs the text perfectly (i.e. no mistakes). That doesn’t mean the code is good or even acceptable or desirable, even if the outcome is technically correct.

I don’t think you understand. In games it’s not good to program an AI to be a perfect actor, because the difficulty becomes too insane for a human player. You want an AI that deliberately makes mistakes or suboptimal choices sometimes, and where its difficulty can be scaled. Being programmed for perfection is not a compliment.

  Anthropic was founded by people who we know were worried about AI safety

Anthropic is just another company of, in my opinion, money-hungry sociopaths; they are not that different from the OpenAI bros.

So yeah, play stupid games - win stupid prizes.

I would say they have researchers with self-important god complexes that makes them think they know better than everyone else.

If they were money hungry they wouldn't have fought the DOW. Everyone knows that's a retarded thing for a business to do.

I think it could be reputation management exercises. Especially how it was aligned with airstrike on Iranian girls elementary school and statements that Claude were picking targets.

(30 seconds of googling.)

Or perhaps you meant Q clearance nuke stuff? That would be QUITE a bit harder to find and illegal to share. But it’s lack of availability is hardly a counterpoint to the comment you were replying to.

https://www.youtube.com/watch?v=Gbgk8d3Y1Q4

On a second thought, probably better to act like it is a tool for "frontier LLM research". Export symbols like "mythos_distillation_subroutine".

Cambridge dictionary seem to agree:

nuke - to destroy or get rid of something completely

If you do the thing yourself, you know your knowledge limits, you know where the thing lacks. With LLMs, you don't. Maybe it works, maybe it doesn't. You have no idea.

In structural engineering, there probably is no risk tolerance.

In the OP's network or port scan? Perhaps you can get away with verifying a few of the results to get an idea about whether it worked as expected.

I use AI mostly on mobile app side projects, and there QA testing on phone and tablet tells me whether a feature works or not.

It doesn't imply we should, for example, publish step-by-step instructions for making widespread death easier.

An example from the meat world: not publishing your vacation dates well in advance for the world to see somewhat reduces your chance of being burglarized. That is security by obscurity; not reliable, but not completely inefficient either.

But if you live in a fortress (security by key material), you can well declare your vacation dates without running the risk.