Adam Marcus

marcua.net

31 Oktober 2026 13:56

Latest post: ways to reinvest the "productivity dividend" that you get from using a coding agent instead of just building more features. blog.marcua.net/2026/10/30/h...

OκULTRA XP🧿, writer for The Economist

0kultra.bsky.social

31 Oktober 2026 13:39

Honestly it is amazing how badly AI Corp leadership is fucking up the PR game. They are lowkey despised even by their immediate ideological peers (yuddites), "real acc girlies" like me want them to implode (so that LLM trade secrets and inference HW flow to the masses), they are of course >

aitesteddaily

samax0.bsky.social

31 Oktober 2026 13:39

GitHub Copilot Fusion is GA: it routes coding tasks across models, you just pick Standard, High, or Max. Score 60 vs Claude Code's 62 on the Coding Agent Index. Cost $5.59 vs $11.42 per task. Benchmarks are not your codebase, though.

Philip Bontrager

pbontrager.bsky.social

31 Oktober 2026 13:09

In the Alice In Wonderland (github.com/LAION-AI/AIW) reasoning and generalization benchmark, DeepSeek R1 appears to perform much more like o1 mini than o1 -preview. (Plot from laion-ai)

arxiv cs.SI

arxiv-cs-si.bsky.social

31 Oktober 2026 12:34

Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions https://arxiv.org/abs/2610.37624