Since OpenAI's math release, GitHub contributors keep driving down the integer-multiplication kappa
xennygrimmato_ · x · 2026-10-16
- A tweet reports that since OpenAI's math model release last week, GitHub contributors have been steadily lowering the kappa for integer multiplication, with big gains so far — an early glimpse of agent-swarm-driven research.
- It promotes Sudo Sunday #1: Auto Research, a LlamaIndex community event in San Francisco, centered on the auto-research-agent tool: point it at a repo with an auto-research.md, press play, and let Codex or Claude Code chew on an unsolved problem; multiple agents can run in parallel, optionally sharing checkpoints.
- Early benchmarks suggest checkpoint sharing doesn't speed things up much. Planned discussion topics: is 5 agents for 1 hour better than 1 agent for 5 hours; should agents pass full checkpoints or summaries; how to verify claimed solutions and review the resulting PRs.
More from coding & agent
- Terminal Night 3 talk slides released, missing demo video to follow — 4310sy · 2026-10-16
- Agent developers are sitting on unused API credits that can exceed actual usage — EnvironmentalDay8864 · 2026-10-16
- Notion's bet on agents: abstract away deployment so users and agents just write code — simonguozirui · 2026-10-16
- Four Gemini 4 Argon agents beat Minecraft's Ender Dragon autonomously in 27 hours — Maxence Faldor · 2026-10-16
- Tencent's Terminal-RSI builds recursively self-improving terminal agents, hitting 19.84% on Terminal-Bench 4.0 — tencent · 2026-10-16
- Tsinghua's AutoSFT uses multi-agent collaboration to automate SFT data research — TsinghuaNLP · 2026-10-16