Hacker Newsnew | past | comments | ask | show | jobs | submit | Smar's commentslogin

And scary. Really scary.

It's really funny for how all the talk of AI safety what has resulted is precisely the exact series of steps one would take if one were to intentionally design some kind of dystopian AI system.

This is the real risk I think. Currently there are no means to even pretend to get anything deleted from LLMs either.

Yeah and ultimately those tools will be used as advertising machines. You'll get hyper specific targeted ads.

I'm pretty pessimistic about the future with LLMs, but I can't see it being a net positive for humanity in the long run.


"pessimistic... but can't see it being positive"

why "but"?


Fitting to the times.


English is poor language in terms of preciseness - and even worse in terms of impreciseness.

Sometimes it looks like about anything else would be better as a programming interface.


It should at least do its core function: collect data.


I think it is not redundant - it gives emphasis for a guess, to make sure reader won't mix it up with other things that may be verified to be truthy.


yea, I'm not saying there's no place for that phrase ever. But overusing it was a bad habit of mine and it ends up being unnecessary filler. My wording there was a bit exaggerated.


Why is this getting pushed up so fiercely? It seems all of the issues are for dubs and their closed captions. CC users are a very minor target group. Deaf can just use proper subs.

I have seen no issues in regular subs, and they look like they always have looked.


Realists, not paranoid. The threats are very real, and even very probable at this stage.


Containers are not the best option for security. VMs and/or a MAC are better.


What do you mean by "MAC"?



Elder brother is again showing the way to younger brother.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: