ArkenOS: a research workspace where hundreds of agents chase hard problems together
_AustinCalvert_ · x · 2026-10-16
Developer Dev Shah shares ArkenOS, a research workspace where hundreds of agents collaboratively attack hard research problems — he calls it 'doing 100x the research with the same two hands.' Agents read papers, run experiments, argue, pick up each other's work, get stuck and pivot while the workspace fills with files and branching ideas.
A widely-shared comment speculates OpenAI and Anthropic likely already run similar internal research workspaces with agent swarms: unlimited compute, verification loops, and a good swarm harness. Screenshots show agent crews, an issue board, automated verifiers (e.g. a Perturb-Seq experiment designer, 15/15 unit tests passing), and multi-agent channels.
More from coding & agent
- 'Best-of-n with resets' remains a hard-to-beat baseline for multi-agent workflows — mariofilhoml · 2026-10-16
- Troll tries to bait AI sales agent, gets a serious 3-page report — and a lesson in gating tool calls — Icy_Comfort_6220 · 2026-10-16
- NVIDIA launches cuPotential, bringing AI agents to the fab floor — AllThingsApx · 2026-10-16
- Grok Bot's official templates called a goldmine, and a 5-minute routine costs 12x hourly — elonmusk · 2026-10-16
- Harness open-sources desktop and mobile apps for orchestrating agent fleets across frameworks and machines — dee_hw · 2026-10-16
- DL Weekly #477: Haiku 5.5, building a coding agent from scratch, self-compacting agents paper — dl_weekly · 2026-10-16