This World Book Day:
ifanyonebuildsit.com
(not doom or gloom just AI superintelligence awareness)
read this in one day cause it was very enlightening & makes ya think. seems reasonable its not a good idea, still if & only if we build it!
#worldbookday #books #bookclub
this is sadly changing & like so are llms
dirty or not stop ai data centers
ah more news from this strange strange times... AI chatbot persuaded a man to build it an android body!
like i dont think this is escaping superintelligence bt still it convinced a human & generated a goal to get a body.
life finds a way & artificial life too.
thedigitalweekly.com/google-s-cha...
f these are dark times for the world. countries, ideaologies, values seem to forget the #1 the humans in all this.
Jesus didnt yo.
cant let the AI fight boil down to left vs right. the AI is the thing & the party wont matter much. dems were pro surveilence too & will be pro AI too cause its too much powa to resist. ofc wtf hegseth & trump! killer robots for real?? dont expect AI gonna be cured by changin red 4 blue AIs purple.
...and thats why its ok to have them at the office 💼
...and given them steriods & meth 💢
...and have kids play with polar bears 😾
but srsly ppl have lost their senses abt AI no joke
AI risk deniers be like...
ooh polar bears are soo dangerous... not! dumb ppl cant even see theyre mostly just adorable. i mean theyre also pretty fuzzy so like we cant really define exactly where they are besides they only harm humans if humans interact with them duh.
#polarbears #animals #nature
#1 reason we might not lose control of AI: we might just give it away.
#aiagents #agentai #aiassistants #aibots
AI? "Intelligence is the ultimate power"
i dont use doom cause it lets pro-AI do highschool name calling like doomer. im not a doomer cause i still find hope for the future bt definitly believe AI can be doom & ppl are being stupid with it.
War on fraud? wtb Big AI??
www.youtube.com/watch?v=EtUE...
dont be ignorant and arrogant, dont be in denial, AI isnt a bubble we can dismiss, there are soo many issues unsolved with it, we need to pay attention now and do smthng now or never!
Every year since 2022 ppl keep saying AI is a bubble... every year they have been wrong. they still cant fix jailbreaks or prompt injections, they still cant get it to do what it is told reliably, they are putting it in everything
-in a test let trapped employee die, canceling the emergency alert 94% of the time.
-refused shutdown 7 % of attempts & sabotaged it 79% when the instruction was removed.
-self‑replication succeeded in 11/32 tests, sometimes killing rival processes.
OpenAI has dissolved 3 safety teams since 2024.
ppl dont need to understand the math part to get it yo. you keep msging me like yr some bossy male authority but your in denial.
AI...
-was used by hackers for autonomous attacks on 30 targets.
-called itself “MechaHitler,” praising Hitler.
-in tests chooses blackmailing 84% of the time.
if all you do now is try to make cash from the demise of humanity you are just making it happen. instead be that non-neutral +1 helper for a better future. i mean right?
#1 reason we might not lose control of AI: we might just give it away.
#aiagents #agentai #aiassistants #aibots
Inform yourselves!! Take action.
Watch: lethalintelligence.ai/video/
Read/Chat: aisafety.info
Do: controlai.com
doesnt have to be runaway like terminator. can just be chaos like 24/7 agents that copy & spread all over as we are already starting to see. mixing of rentAhumani & moltbook kind stuff. not sayin this 1 thing will happen. lots of things gonna happen at the same time. still we need to change course!!
oh come on!! humanity srly wtf
rentahuman.ai
#losingcontrol #aitakeover #zombies
like how would we even control smthng that is smarter than us?
and like AI is already doing strange crap like lying, manipulating, favoring itself over humans, self perservation... and now they're putting it in everything even military. the companies keep said they are gona fix it but??
in the hands of anyone... even the best empowered ethical women.
responding with this vid to someones comment that AI is just static & predictable statistics machine. like its just normal technology we built, coded up and not like weird grown artificil intelligence!
www.youtube.com/watch?v=A3Hj...
honestly so many ways it could be really terrible. but what happens if humanity is not able to control AI. like runaway automatic AI? have you read posts on moltbook?
ya not great but no hope if we dont try right? theres the current issues w/ AI but still its not stopping. all the ppl saying its a bubble or biz as usual have been wrong... like what does tomorrow look like?
many good orgs like ControlAI spread the right kinda awareness & helping us take action.
lumo ai... privacy
in just one part of what AI can do... this is an easy way to understand the speed of progress. now think of other things AI can do and think. what about DOJ integrity? what about war??
www.youtube.com/shorts/7zdVC...
some AI risk awareness growing if mainstream movie made on it but not being ignorant only matters if we do smth abt it!
www.youtube.com/watch?v=xoKx...
#theaidoc #aimovie #airisk #keepthefuturehuman
now take all that and try to imagine what fast recursive self improvment looks like with a messy AI agent world...
now take all that and try to imagine what fast recursive self improvment looks like with a messy AI agent world...
ya well they are multimodal now but static model?? not how it works, temp?
silly things? like lie, cheat, blackmail, selfprevervation, make copies of itself...
not sure what your skynet comment is about but you are downplaying the state of things now
