Thread 106181625

We are almost there!

>>106181625 (OP)
holy FVCK… ASI in 2026…

>>106181691
probably

>>106181625 (OP)
TWO MORE WEEKS

how does it score on FINGER-AGI-V1 though?

>>106182285
>I NEED $100 BILLION MORE TO FIX THIS!!!!!

>>106182285
the issue here is how confident it is. it really breaks the illusion

>>106181895
for sure this time!

these people fucking vote, never forget that
can't even see through an obvious marketing scam

and these people are above average intellect compared to the actual morons
democracy is the stupidest thing to have ever existed, at least Jews are in charge, ironically it would be worse if they weren't

>>106182326
>these people fucking vote, never forget that
>can't even see through an obvious marketing scam
investors?

>>106181625 (OP)
Oh shit guys, it's almost intelligent, there's just no way they trained it specifically on these kinds of problems now, nope

>>106182285
>how many Rs in strawberry
>you sure about that bro?
>chatgpts face when

>>106182911
There are three R's in the word "strawberry".

>>106181691
artificial shithead intelligence?

>>106182306
Yeah but that's fundamental to how LLMs work.
They have to bullshit and they have to seem confident about it because they don't know when they are right or wrong because they don't work that way,
BECAUSE LLMS ARE TOKEN-AT-A-TIME AUTOCOMPLETE AND NOTHING MORE

>>106181691
Artificial Mega Intelligens 2027 confirmed

>>106183286
raspberry isn't hard coded on the correct response like strawberry. I would bet it still fucks it up if it weren't for that

>>106181625 (OP)
Nah another meme benchmark with still no real intelligence to be found.

>>106181625 (OP)
>$10 per prompt
wow that is impressive

>>106183286
Horizon Alpha is GPT-5 on Openrouter.

>>106181625 (OP)
AGI in two more weeks, Goy, pay me One Trillion Dollahs.

>>106181625 (OP)
>Another luddite thread
Codetroons coping rn