Caleb Writes Code

▶

▶ 8.4K ❤ 0 💬 0

Can we appreciate the breakthrough that OpenAI made in Navier-Stokes without all the drama behind it? OpenAI took a similar approach by taking C/D counter-example by blowing up Euler's equation without forced which is something that Buckmaster and Alpoge was not able to do, though with enough time, they could have gotten there. While we are still finding out more details on denied allegation of OpenAI cheating, it's important to understand first how Navier-Stokes actually works, and second understanding the approach taken by OpenAI that led to the breakthrough.. which would likely need another 1-2 years for the solution to be accepted or rejected by the institution. #math #mathematics #newton #newtonian #physics #openai #ai #artificialintelligence #largelanguagemodels #machinelearning #deeplearning #graphicscard #LLMs #inference #datacenter #gpu #explained #aibubble #anthropic #harnessengineering #compute #research #spacex #spacexai #agi #xai #llm

▶

▶ 21.7K ❤ 0 💬 0

Substack: https://substack.com/home/post/p-215679920 DeepSeek-V4.1-Flash is beautifully crafted model that introduced so many bold takes on model architecture. With CED, the model only activates 50% of the model during prefill which saves huge compute overhead, and CSA2 on top allows KV Cache reduction without sacrificing quality. mHC optimization they added in kernels by simply rearranging the algebra on the residual operation also helps save communication overhead between SM and HBM. Every inch of the model was built with intention and I can see it reading through the entire research paper. Great work @deepseek_ai team in constructing this, but more importantly, sharing the knowledge to the world. This is invaluable. #deepseek #artificialintelligence #largelanguagemodels #machinelearning #deeplearning #graphicscard #LLMs #inference #datacenter #gpu #explained #aibubble #anthropic #harnessengineering #compute #research #spacex #spacexai #agi #xai #llm

▶

▶ 17.1K ❤ 0 💬 0

Jev, a new type of model that is taking over how we think about getting information out of the model. #jev #ai #tech #artificialintelligence #largelanguagemodels #machinelearning #deeplearning #graphicscard #LLMs #inference #datacenter #gpu #explained #aibubble #anthropic #harnessengineering #compute #research #spacex #spacexai #agi #xai #llm