𝙻𝙰

[𝙻𝙰 @LA1986]

@IbrahimDagher20

[𝙻𝙰 @LA1986]

@var_epsilon

[𝙻𝙰 @LA1986]

@patloeber @jeffecom They’re honestly decent models but there are so many very good models available to choose from that decent doesn’t cut it.

[𝙻𝙰 @LA1986]

@BBCWorld So 375 billion then

[𝙻𝙰 @LA1986]

@APompliano Of course the US can compete in AI with China on performance of a specific model, but never on price.

[𝙻𝙰 @LA1986]

@Austen Maximum cope.

[𝙻𝙰 @LA1986]

@suchenzang

[𝙻𝙰 @LA1986]

@teocns @togethercompute

[𝙻𝙰 @LA1986]

@zeroxkyle

[𝙻𝙰 @LA1986]

@cnnbrk I’d be surprised if Trump doesn’t intervene to stop his human trafficking bros from being deported.

[𝙻𝙰 @LA1986]

@deanwball Dude, we get it, you work for a closed-source frontier lab and you were hired to spread the idea that open-source AI is bad.

[𝙻𝙰 @LA1986]

@doomslide Washington really produces the very worst type of people. The fact that they’re trying to bring back Communism as the boogeymen in 2026 is comical. China is more capitalist than Communist these days.

[𝙻𝙰 @LA1986]

@NicoHezel @Alibaba_Qwen Small models are useless

[𝙻𝙰 @LA1986]

@Alibaba_Qwen

[𝙻𝙰 @LA1986]

@deanwball Lmao OpenAI trying to tell you that Open AI is bad, the irony.

[𝙻𝙰 @LA1986]

@Glorifundel @RasmusJarlov

[𝙻𝙰 @LA1986]

@zxytim You guys cooked for sure, what are the chances you could significantly reduce the inference costs in follow-up revisions of K3?

[𝙻𝙰 @LA1986]

@OmedVibeCodes DeepSeek about to drop Kimi K3 level model for dirt cheap and put everyone else out of business.

[𝙻𝙰 @LA1986]

@ID_AA_Carmack Ha yeah true, pretty obvious really.

[𝙻𝙰 @LA1986]

@doodlestein @sundarpichai what gives?