Stories by GreenGames

256K context with 72MiB of KV cache on the GPU

A 35B MoE on a 16 GB GPU, without the offload tax

PFlash: 10x prefill speedup over llama.cpp at 128K on a RTX 3090

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

Show HN: OS Megakernel that match M5 Max Tok/w at 2x the Throughput on RTX 3090

Cua (YC X25) is hiring an engineer

App-Use, Control Individual Applications with CUA Agents

Show HN: Lumier – Run macOS VMs in a Docker

Microsoft is reportedly about to lay off 3% of its workforce

Polaris is giving free GPUs/CPUs for everyone

Improvements in reasoning AI models may slow down soon, analysis finds

Microsoft and OpenAI are renegotiating their partnership

Apple developing new chips for smart glasses, Macs, and more

An AI Code Assistant into Your IDE