I really hope this ever becomes realbut the thing is, a wafer is a wafer, it has finite areawe can tweak the architecture, but it'll be…
Xi likes the United States and is a serious person who thinks a lot about the value of technological revolutions China had missed out on. So…
CoolGrok: Grok Build and Grok 4.5 top user model preference
RT Dr Jordan B PetersonWe think of this evil as Marxism. In reality, Marxism is nothing more or less than the current instantiation of a much…
RT Longhui YuDeepSWE is one of those benchmarks where model differences really show. Excited to see Kimi go from K27-code's 31 points to a 2x leap…
Grok 4.5 is arguably #1 when taking speed and cost into accountX Freeze: Grok 4.5’s efficiency is ridiculousOn Artificial Analysis, it costs just $0.31 per Intelligence…
RT ThePersistenceOlive Garden implemented the SAVE America Act more quickly than the US Senate.
Ollama 0.32.1 includes significant improvements to Gemma 4's tool calling, making it much more reliable in coding agents!To try Gemma 4 26B with Pi, run:ollama launch…
RT Lewis 🇺🇸Yes! Not even “would’ve,” but DID NOT: He was here in America! They clearly did NOT offer a better deal, it was clearly NOT…
GLM is much stronger than V4 in cyber and neither amounts to muchAI Security Institute: On our cyber range "The Last Ones", GLM-5.2 matches Opus 4.5,…
jensen is really twitchinghe doesn't want to discuss his misadventures in ChinaMackenzie Hawkins: A senior US official, three days ago: a "trivial" number of Nvidia H200…
So we know that a 1e25ish pretrain with mediocre post-training, the scale of V4 or others in this bracket, is enough for 36.5% on ARC-AGI-2. GLM-5.2…
Inkling’s humility at launch may have made it underrated. The previewed small model will wake a lot of people up (if it comes soon). Look at…
RT MaxThe beginning of the Megacharger networkBuild-out is focused on the most important Semi routes and depots first. Follow our progress here: https://www.tesla.com/findus?bounds=40.088148238001295%2C-72.5347021341828%2C25.057723630856046%2C-110.0859716654328&filters=tesla_exclusive_megachargers%2Ccoming_soon_megacha
Inkling is now running inside @opencode through Together AIZain: just added Inkling to OpenCode🔥awesome model from @thinkymachines running on @togethercompute!
If you don't have drone fighters, you're ngmiI don't really know whom this matters for, though. Russia, Canada, Europeans probably. Countries without serious aviation are cooked…
It should be the golden age of porting. Maybe iPad will finally get some good games.Kahris: I used Codex 5.6 Sol to get RollerCoaster Tycoon 2…
RT Aston Martin Aramco F1 TeamAMR Network: Portraits with @cohere.The volume of data behind every race weekend is vast, and how it's processed matters as much…
Frontier open models are accelerationist on ai diffusion and deflationary on the valuations of the biggest tech companies with large capex. IMO that tends to be…
Teaming up with http://You.com for their Agentic Hackathon at the AWS Builder Loft in SF, July 24th. One day to build agents on live web data…
Will be bigger than you thinkperhaps labs needed to reach a certain levelZephyr: @PatrickMoorhead So, a lot of Mercor/Surge type companies have popped up in China…
RT PeléThe goal Pelé called the most beautiful of his life was never filmed. Aug 2, 1959, vs Juventus, on Rua Javari. With @Google, we rebuilt…
RT hardmaruReal brains follow Dale's principle: a neuron can either excite its neighbors or suppress them, but never both. Standard deep learning ignores this and uses…
Real brains follow Dale's principle: a neuron can either excite its neighbors or suppress them, but never both. Standard deep learning ignores this and uses backpropagation.In…