I vibecoded this neural network visualization for my students and open sourced it.
It shows a simple MLP trained on MNIST handwritten digits at several training steps. The visualization is using @threejs and it comes with training code in @PyTorch .
Link + repo 👇
An archaeologist friend of mine wished he had a certain software for 10 years. I vibecoded it for him that evening over a bottle of wine.🍷
He never even used ChatGPT before, and his mind was blown. 🤯
Link, repo, and story below.👇
The perfect spam filter for AI automated applications. Anyone who claims to fulfill this is a fraud and everyone who understands this will rather quote related work in other areas of unconventional computing.
DGX Spark GPT OSS 120B: 11.65 tok/sec
M3 max GPT OSS 120B: 41.71 tok/sec
This is very bad. Can't be explained from bad ollama performance and slower memory speed (273GB/s vs 400GB/s) alone.
🚀 SGLang In-Depth Review of the NVIDIA DGX Spark is LIVE!
Thanks to @nvidia’s early access program, SGLang makes its first ever appearance in a consumer product, the brand-new DGX Spark.
The DGX Spark’s 128GB Unified Memory and Blackwell architecture set a new standard for
It's a bit more than that. Finance companies are under severe pressure in China (as it is seen as unproductive labor) and quite some are looking for pivot opportunities. Wouldn't wonder if they fully pivot to AI and make quant their side project.
Not if you redesign the system every 4 weeks. This is rapid iterating war production and not a consumer product with several year lifespan. It’s exactly this out-of-box manufacturing that gives Ukraine the edge in the drone wars.