I have very little idea of what K3 might deliver what hints we have are pretty intimidating.
I have very little idea of what K3 might deliverwhat hints we have are pretty intimidating.Elaina: Kimi K3 in July.
Latest
best safety practices for 2026: – Russian language pack (to not get targeted by ruskie cyberattacks) – PDT as time zone other tips?
best safety practices for 2026: - Russian language pack (to not get targeted by ruskie cyberattacks)- PDT as time zoneother tips?AI超元域: 🚀劲爆消息！Claude Code 被曝疑似内置“隐藏后门”，专门检测中国用户。Claude封号原因终于找到了！！！据 Reddit 爆料：从…
I like how Tyler just accepts that this guy's name is 'Clavicular'
I like how Tyler just accepts that this guy's name is 'Clavicular'Richard Hanania: Tyler Cowen says Clavicular is the future.
…
…tphuang: Oh boy, UBTech's U1 bionic robots now on sale. Low end goes for 119.8k RMB & high end version are 880k & 990k RMB.Uses RK3588…
The token efficiency really is striking do we have a more charitable explanation? Cursor bros Just That Good at RL?
The token efficiency really is strikingdo we have a more charitable explanation? Cursor bros Just That Good at RL?Will: I think i understand why @cursor_ai doesn't…
The plot thickens! I guess Tianhang can boast of his RL contributions very legitimately. PPO back then was a really high-risk, high-effort choice. I a…
The plot thickens! I guess Tianhang can boast of his RL contributions very legitimately. PPO back then was a really high-risk, high-effort choice. I apologize.Zheng Yuan:…
Huawei Pangu's 92B 6B active MoE open sourced Big one coming in July They're basically trying to do Nvidia's thing with Nemotron
Huawei Pangu's 92B 6B active MoE open sourcedBig one coming in JulyThey're basically trying to do Nvidia's thing with Nemotron智东西China AI News: 🚨TODAY: #Huawei open-sources OpenPangu-2.0-Flash#OpenPangu…
We're about to start seeing feats that are less heroic and more scary. By the fourth week of July 2026, we are likely to see: – DeepSeek-V4-official -…
We're about to start seeing feats that are less heroic and more scary.By the fourth week of July 2026, we are likely to see:- DeepSeek-V4-official- Huawei…
Underappreciated detail of Meituan claims: as far as I can tell, Yunfan is correct, they run CloudMatrix 384 pods in "910B mode". So it's not 50K GPUs…
Underappreciated detail of Meituan claims: as far as I can tell, Yunfan is correct, they run CloudMatrix 384 pods in "910B mode". So it's not 50K…
> Another shocking data point is DSpark's TPOT, which is only 2.9-5.2ms, indicating that the built-in neural network layer in DSpark runs exceptionall…
> Another shocking data point is DSpark's TPOT, which is only 2.9-5.2ms, indicating that the built-in neural network layer in DSpark runs exceptionally fast. The latency…