Show HN: Ran Qwen 3.8 27B autonomously(ish) for 3 weeks on single 3090huggingface.co·2 pts·iamskeole·0
Show HN: Don't trust me bro: fixing GPT-OSS (3.49B tokens, 1k GPU hours, 1x3090)github.com·2 pts·iamskeole·0
Frontier AI Models Exhibit Sophisticated Reasoning in Simulated Nuclear Crisesarxiv.org·3 pts·iamskeole·0