281–290 of 450 posts
Earlier quoted context omitted.
Orwell's idea of "wrongthink" is more relevant than ever.
Orwell also imagined a future where Speakwrites made everyone unable to write by hand, making their thoughts dull and tanged as a long unprepared monologue tends to be. I don't think he's anywhere on your side.
Earlier quoted context omitted.
Every AI would refuse the prompt. I was banned for researching Nordic assisted death and asking which drug exactly they administered (and what quantity). Claude refused, alerted Anthropic, and I was banned a couple days later. Thankfully the appeal form worked, but by then I was using a different Claude account praying they didn’t ban me again. There’s an uncensored model floating around that you can run locally with…
The ease by which you can get banned worries me. It seems unavoidable that soon AI will manage its own suspicion level, provide feedback on it, and when high enough it will call the authorities.. because that's what people do. Banning doesn't cut it, like you can't deny internet access. Soon this will spiral out of control and AI (Palantir) will have to run the response and the parallel AI state erects itself. A citi…
I can totally get the "don't tell people to kill themselves" aspect of these models, but certain parts of the Internet have been telling people that for decades.
Certainly the models should be trained/tuned to avoid conversations like that wherever possible and redirect people to get the help they need...but that's exactly the problem; doing that is MORE than what the state and the strangers surrounding a person would do. That's the problem, a mental health crisis that is ignored, particularly in men.
Earlier quoted context omitted.
Every AI would refuse the prompt. I was banned for researching Nordic assisted death and asking which drug exactly they administered (and what quantity). Claude refused, alerted Anthropic, and I was banned a couple days later. Thankfully the appeal form worked, but by then I was using a different Claude account praying they didn’t ban me again. There’s an uncensored model floating around that you can run locally with…
>researching Nordic assisted death and asking which drug exactly they administered (and what quantity). Odd ban. The countries with legally passed instances use similar drugs/processes — you could research about those, wiki, google, actual websites, dignitas et al. No ban needed. Was the ban reason spelt out? Was it your wording?
Humans aren't like that lmao. We're reactionary, tribal animals.
Earlier quoted context omitted.
It has nothing to do with human decisions. Bans always come later because it's a best way to make sure most people who get banned would never know what are they banned for exactly and what threshold there is, The same tactics used in game development against cheaters. If it would ban you right after prompt you'll know how to avoid getting banned. Obviously that didn't worked for you because you wasn't doing multiple…
Except in the case of video games it in practice means that cheaters get to terrorize your playerbase without being banned. This tactic is the kind of decision that is made by people staring at metrics all day without considering what's going on in reality.
One of my friends in high school used to cheat on a popular video game. The fact ban waves would occur about once a week to once a month meant whenever his accounts got banned, he never knew why exactly and wasn't able to stop it the next time.
Of course, if ban waves are too long apart then yeah you're just letting a known cheater wreak havoc on the playerbase.
Earlier quoted context omitted.
In the near future, the only way to tell if someone is human will be to have them say a slur. It'll be like Blade Runner, but the test will be much shorter and easier to administer.
It is already a way to filter some people from others, in particular adults from children. https://www.paulgraham.com/say.html
In case anyone else wanted to get to a demo quickly: https://xcancel.com/MaximeRivest/status/2073544461473169432#... Original link for those who want it: https://x.com/MaximeRivest/status/2073544461473169432
Earlier quoted context omitted.
> Cars are literally the number one killer of children 0-14 in the United States. More than cancer, more than guns, more than the next 7 reasons combined Cars are heavily regulated. Require licenses to use. Entail massive losses of freedoms when used, e.g. you can be randomly breathalyzed or whatnot. Cars are dangerous. Comparing anything to a car (or, per Altman and now Dario, to a nuke) means it should be tightly r…
Essentially no one in the USA fails to get a license if they try for one. There's no ongoing competence testing. Washington State recently failed to advance legislation to make the fourth DUI have a jail sentence because it would be too expensive imprisoning that many people . > Cars are dangerous. You can kill yourself with a kitchen knife in your own kitchen. Should be "heavily regulate" that too? How about we have…
About that alignment issue ...
Earlier quoted context omitted.
> if by magic The problem is the means not the ends. This magic doesn’t exist. In practice is authoritarian surveillance. If we could kiss a mushroom and bias the universe’s dice such that crime just…wouldn’t happen, yes, that would be good, though it would also open up a plot hole of consequences we, in the real world, don’t need to worry about in general.
It won't be good, as the definition of what's a crime and what is not changes due to political reasons (read corruption).