cannot shoot yourself
would you save a drowning child if it meant you had to get your clothes wet
I am sorry for what I am about to do with the ceiling
And then, word will filter down from the O5s, through the many levels of our bureaucracy. And those who were late to work anns unethical will be given reprimands which will be noted on their permanent record. Or their pay will be cut, or they will be demoted, or they will be transferred to another project.
Or they will be shot for crimes against humanity
shockingly, building a mind from scratch is hard. you canât just say âwell we put all the data in a big bowl and swirled it around a bit and presumably if we find the right way to swirl it itâll naturally produce an artificial reddit userâ, itâs something thatâll take, you know, hundreds of fucking years
I see no need to fight over anything
The fact that weapons development is necessary sucks absolute balls but it is an unfortunate reality of the world we live in that far too often might makes right
And the best we can do given the scale of humanity as a whole is hope that we never have to actually use offensive weapons while making sure our defenses are solid (and, mind you, most modern defense is basically offense - the most reliable way to intercept a missile is with another missile built for the purpose)
tl;dr war sucks and having worked at a defense contractor through the start of the war in Ukraine believe me when I say that nobody believes in that more than the people working to build the munitions they are fought with - after all, who else knows better what these things can do?
1 Like
why are you shooting the ceiling
perhaps fifteen thousand
Gosh, I wish we could find the methodology behind swirling the artificial reddit user
Testing a game theory
the fact that you felt bad about contributing to peopleâs murders doesnât absolve you of the fucking guilt
fucking scum.
youâre conflating a number of different claims and treating them as broadly equivalent when they arenât
âno one at OpenAI/other AI companies has tried to understand their AIsâ - blatantly false
âthe only way theyâve tried to understand their AIs is by having the AI explain the AIâ - also blatantly false
âno one at OpenAI has made more than minimal progress at actually understanding their AIsâ - clearly true imo, I donât think most people at OpenAI/etc. would even dispute this
âOpenAI is trying less harder than they Ought To to understand their AIsâ - probably true, but doesnât prove or disprove any of your other claims
âan LLM could never tell you that the entire basis for the field is inherently flawedâ - technically false in the sense that I could surely get ChatGPT to write an argument for it, if Iâm not being pedantic then itâs just not very relevant? like if youâre trying to understand the inner workings of the AI, you can try to do that whether or not the entire basis of the field is inherently flawed
âLLMs are not very useful for ?useful real world things?â - entirely different claim with little relationship to the previous ones
âOpenAI is badâ - entirely different claim with little relationship to the previous ones
I feel exactly zero guilt over working at a company providing munitions to Ukraine and if you think I should thatâs a lol from me dawg
1 Like
literal definition of a country minding its own business when an aggressor walked in and started shooting
2 Likes
team Polish infighting
1 Like
Was I not naked at the time?
2 Likes
considering the actions the russian army has taken this is not a hill i would be willing to die on personally
2 Likes
hi what the fuck does this mean
2 Likes
one soldier recorded himself doing things i obviously cannot say on this forum to a literal toddler i have zero empathy
1 Like