Dale Ren

uchicago
exploring and exploiting

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[Nathan Lambert @natolambert]

TLDR: An openai model, during evaluation on a cyber benchmark, exploited a public zero day bug, escaped sandboxing in openai's infra, and got into the internal huggingface infra via an exploit (through a public dataset service) all in the attempt to solve a benchmark problem.

[Dale Ren @DaleRen2]

@Polymarket fork found in the kitchen

[Dale Ren @DaleRen2]

why the fuck are we buying the argentinian peso

[Dale Ren @DaleRen2]

@esharwood_ i think you're right

[Dale Ren @DaleRen2]

the question is: do i not understand the proof or is gpt hallucinating?

[Dale Ren @DaleRen2]

@esharwood_ 😭

[Dale Ren @DaleRen2]

early bird gets the worm type shi

[Dale Ren @DaleRen2]

@Tetris_Official i need tetris friends back

[Dale Ren @DaleRen2]

@glenn_thinks ayo

[Dale Ren @DaleRen2]

@glenn_thinks i read this 6 times and have no idea what you’re saying

[Dale Ren @DaleRen2]

@glenn_thinks we ballin

[Dale Ren @DaleRen2]

@RGIII cost me in fantasy 🤦‍♂️

[Dale Ren @DaleRen2]

@glenn_thinks team > idea

[Dale Ren @DaleRen2]

@glenn_thinks ligma balls

[Dale Ren @DaleRen2]

@glenn_thinks -kyrie

[Dale Ren @DaleRen2]

@glenn_thinks interacting w bitches

[Hank Green @hankgreen]

Might fuck around and try to substantially lower the cost of higher education in America.

[Dale Ren @DaleRen2]

@glennren lmaoo

[Dale Ren @DaleRen2]

@Momathan1 never watched it bruv