Image created with gemini-3.1-flash-image-preview with claude-sonnet-4-5. Image prompt: Vintage 1990s screen-printed t-shirt graphic on worn mustard-yellow cotton fabric, deep red ink only, showing cartoon deep-sea diver in brass helmet kneeling on ocean floor with magnifying glass examining treasure, simple seaweed and bubbles, bold text DEEPSEEK arched across top in chunky retro lettering, faded print texture with slight fabric stains, humorous local novelty shirt style
DeepSeek MoE training efficiency has been completely commoditized (though it’s startling how much of that is reimplementation of DeepSeek’s own Open Source Week releases 1 year ago, updated for Blackwell). Very cool
https://x.com/teortaxesTex/status/2031263831595335702
🧵 1/4 Still waiting for DeepSeek-V4? We (@Zai_org) made DSA 1.8× faster with minimal code change — and it’s ready to deliver real inference gains on GLM-5. IndexCache removes 50% of indexer computations in DeepSeek Sparse Attention with virtually zero quality loss. On GLM-5
https://x.com/realYushiBai/status/2032299919999189107
Our next kernel competition is now open for submissions! A $1.1M cash prize competition sponsored by AMD on optimizing DeepSeek-R1-0528, GPT-OSS-120B on MI355X Registration:
https://x.com/GPU_MODE/status/2029974019018244223
Leave a Reply