mji

Recent Submissions

1. 1. ▲ Snapcompact: SoTA compaction – instant, local, free. Pick 3 (stencil.so) 2 points · 3 days ago · 0 comments
2. 2. ▲ The Future of Meta Superintelligence: A 1 Year Progress Update (newsletter.semianalysis.com) 4 points · 25 days ago · 1 comment
3. 3. ▲ DeepSeek-V4 on Day 0: From Fast Inference to Verified RL with SGLang and Miles (lmsys.org) 80 points · 3 months ago · 10 comments
4. 4. ▲ TorchTPU: Running PyTorch Natively on TPUs at Google Scale (developers.googleblog.com) 206 points · 3 months ago · 18 comments
5. 5. ▲ RvLLM: High-performance LLM inference in Rust (github.com) 3 points · 4 months ago · 0 comments
6. 6. ▲ My chief of staff, Claude Code (twitter.com) 3 points · 4 months ago · 6 comments
7. 7. ▲ Native KV Cache Offloading to Any Filesystem with LLM-D (llm-d.ai) 2 points · 5 months ago · 0 comments
8. 8. ▲ Mooncake Joins PyTorch Ecosystem (pytorch.org) 1 point · 5 months ago · 0 comments
9. 9. ▲ The K-Shaped Future of Software Engineering (twitter.com) 1 point · 6 months ago · 0 comments
10. 10. ▲ How Prompt Caching Works – Paged Attention and Automatic Prefix Caching (sankalp.bearblog.dev) 6 points · 8 months ago · 0 comments