Microsoft Research has open-sourced Agensh, a multi-agent harness that drops the central orchestrator most agent frameworks rely on. Workers pick their own sub-tasks and coordinate via a shared Git repo, a chat channel and a findings board. On ProgramBench, the mean test pass rate rose from 19.31% with one agent to 28.78% with 128, and bigger teams hit milestones sooner. https://www.marktechpost.com/2026/10/24/microsoft-agensh-multi-agent-harness-1024-agents/ #AIagent #AI
Related
When games stop being games #negativepid #digitalInvestigations #OSINT #cybersecurity #AI #tech #onlineInvestigations #r...
When games stop being games #negativepid #digitalInvestigations #OSINT #cybersecurity #AI #tech #onlineInvestigations #robotics #cyberpsychology #cybercrime https://negativepid.blo...
While Republicans and some treacherous Democrats in the Senate suck off the AI fascists we just defeated so we could eve...
While Republicans and some treacherous Democrats in the Senate suck off the AI fascists we just defeated so we could even have elections, the AFL-CIO is drafting comprehensive new ...
And two it refuses to answer: sentiment and reach. The data isn't there, so the tool isn't either. A number that looks r...
And two it refuses to answer: sentiment and reach. The data isn't there, so the tool isn't either. A number that looks right but isn't is worse than no number.15 of the 36 are new ...