HNHacker News
TopNewBestAskShowJobs

NicoConstant

80 karma · joined March 13, 2026

submissionscomments

Real-time LLM Inference on Standard GPUs: 3k tokens/s per request

blog.kog.ai·219 pts·NicoConstant·
97

Kog AI – Building a Real-Time Inference Stack on AMD Instinct GPUs [video]

youtube.com·8 pts·NicoConstant·
0