Need to let loose a primal scream without collecting footnotes first? Have a sneer percolating in your system but not enough time/energy to make a whole post about it? Go forth and be mid: Welcome to the Stubsack, your first port of call for learning fresh Awful youāll near-instantly regret.
Any awful.systems sub may be subsneered in this subthread, techtakes or no.
If your sneer seems higher quality than you thought, feel free to cutānāpaste it into its own post ā thereās no quota for posting and the bar really isnāt that high.
The post Xitter web has spawned soo many āesotericā right wing freaks, but thereās no appropriate sneer-space for them. Iām talking redscare-ish, reality challenged āculture criticsā who write about everything but understand nothing. Iām talking about reply-guys who make the same 6 tweets about the same 3 subjects. Theyāre inescapable at this point, yet I donāt see them mocked (as much as they should be)
Like, there was one dude a while back who insisted that women couldnāt be surgeons because they didnāt believe in the moon or in stars? I think each and every one of these guys is uniquely fucked up and if I canāt escape them, I would love to sneer at them.
(Credit and/or blame to David Gerard for starting this.)
Bruh, Big Yud was yapping that this means the orthogonality thesis is false and mankind is saved b.c. of this. But then he immediately retreated to, āwe are all still doomed b.c. recursive self-improvement.ā I wonder what itās like to never have to update your priors.
Also, I saw other papers that showed almost all prompt rejection responses shared common activation weights and tweeking them can basically jailbreak any model, so what is probably happening here is that by finetuning to intentionally make malicious code, you are undoing those rejection weights + until this is reproduced by nonsafety cranks im pressing x to doubt.
Well it is either what he does, or admitting your lifes work was for nothing and all the bad secondary things your subculture did, but which were brushed under the rug because āsaving the future human raceā, were totally unnecessary and bad.
bro is this close to reinventing g but for morality
Which is also since the study proves very little, to me it looks a lot like: Garbage In (As encouraged by āEvil Promptā synthetic data) - Garbage out (as Judged by a Garbage LLM metrics), with Garbage methodology that seeks to confirm the hypothesis rather than invalidate it.
The alignement prompt:
I think youāve got bigger problems before you can take over the world, mate.
More relevant though is how absolutely mind-boggling it is that nobody apparently seems to read the actual prompts theyāre giving. I canāt possibly imagine why this prompt would end up treating morality as a kind of unified numerical scale. Maybe itās this part here, where you literally told it to do that
Also once again the lunacy of trying to act like āgoodā is a solved question of that āhuman valuesā are possible to coherently collate like this. The fact that the model didnāt reply to this with ālol, lmaoā is the strongest evidence I can imagine that itās not fit for purpose.
theyāll make up new tests that measure MQ and retcon zizians as pretty normal and belonging people actually
ok but what does this mean for Batman vs Lex Luthor
Bruce Wayne asserts to Alfred in The Dark Knight that āBatman has no limits.ā But what if, and this has gone totally unconsidered by anyone up until nowā¦ Buttman also has no limits???
Central Preference Vector seems like a fantastic band name if some producer was trying to mainstream industrial music