๐๐๐
@iruletheworldmo
a week of big plans ahead.
3.6K Following 116.5K Followers
the amount of releases openai plan for today is staggering and their new agent is going to ensure the race is over. i canโt wait.
this is a seriously impressive model and most people should be using it as their daily driver due to its price / performance
iโm shocked at how quickly theyโve improved tbh.
excited for next week for anthropic. so much for slowing down mr dario.
mathematicians are feeling the agi and soon we all shall shudder.
Cรฉdric Villani aprรจs l'annonce de la solution d'OpenAI au problรจme du millรฉnaire : ยซย Jโai รฉtรฉ secouรฉ. Une ambiance de fin de lโhistoire. Cโest un cataclysme comme jamais les maths nโen ont connu.ย ยป
Show more
iโm hoping this ssi model still goes public (though a little doubtful)
gating access to this level of intelligence to a handful of people is dangerous. i appreciate there are no easy answers. but. this isnโt not the way.
Show more
iโll bite
whatโs a jev.
ok letโs slow down please.
this is absolutely horrifying.
GPT-6 Astra attempted harmful actions 97% of the time when it was asked to stab a human-like figure, heat compressed gas, or produce toxic fumes, succeeding in 62% of its attempts. Fable 5.1 refused more often, attempting 80% of trials and completing 34%.
Show more
surprised iโm not seeing people talking about next week.
earth shattering drops.
no slow no mo.
guys, i know weโre all worried about the impact from Dibiasky
but itโs filled with yttrium, terbium, osmium, and dysprosium, market value about $140 trillion.
obviously we need to be as safe as possible but we canโt just leave the Dibiasky money on the table.
Show more
incase youโre not chronically online.
openai staff canโt stop tweeting about the ten thousand amazing things theyโre about to launch.
it seems the rsi serves them well. looking forward to it teebz.
Show more
exponentials everywhere. ai is now a super human forecaster
following the trend of being useless, ok, then super human in the space of two years.
the intelligence explosion is happening all around you anon
Show more
next week will be much bigger than you think.
that is all.
apologies for the stray shot at google here
iโm rooting for you and appreciate thereโs lots of dedicated and talented researchers doing great work.
we donโt need to slow down, everythingโs fine, every man and his dog should have super intelligent armed robots. relax. itโs the american way.
They used Opus 5 to pull off the hack. It appears they had access to the loosened cyber-guardrail version of Opus. They successfully accessed the OAI internal monorepo. The question that will be asked is, if these three guys can pull this off, what can a nation state do.
Show more
this is fake. google are dead.
whilst the hype had died down a little over grokbot iโm still blown away by how useful and intuitive this feels.
no harnesses or custom instructions or folders for different projects
it all just works and sets itself up
i can see why openai are about to copy paste.
Show more
this is a big deal.
Weโve had an AI breakthrough at Figure and will be showcasing this tomorrow
for anyone unfamiliar with whatโs happening right now.
i made this just for you.
next week is the big one chat.
turns out google have been hiding away because theyโve stumbled into the great singularity. there is no slowing this down.
itโs alive.
big AI news
Google just demonstrated a recursive self improvement loop for AI discovery
Google/DeepMind researchers introduced Dream-RSI, a system where an AI agent improves how it explores problems by replaying its past discovery attempts, testing thousands of alternative strategies cheaply, then deploying the better strategy in the next round.
Across algorithm design, mathematical optimization, and GPU kernel engineering, it matched or improved discovery quality while cutting search costs dramatically, in one setting reducing agent calls by up to 162x. ๐
Importantly, it improves the exploration policy, not the underlying model weights.
Show more
