1I ran forensics on closed models and discovered no one is using dense attention (opens in new tab)(blog.0xmmo.co)2mmoustafa10d ago0
2Fastest small LLM at 1 KB context is the slowest at 1 MB (opens in new tab)(blog.0xmmo.co)1mmoustafa11d ago0
3Show HN: An AI assistant you can text via Apple satellite messaging (opens in new tab)(olly.bot)1mmoustafa4mo ago1
4Preventing agent doom loops (with reasoning traces) (opens in new tab)(0xmmo.notion.site)1mmoustafa4mo ago0
5Preventing agent doom loops with Reasoning Traces (opens in new tab)(0xmmo.notion.site)1mmoustafa5mo ago0
7Trump's iconic writing style partly due to Truth Social UIKeyboardType choice (opens in new tab)(twitter.com)6mmoustafa7mo ago1
