Startup humans& launches Ampersand, a model trained for collaboration instead of autonomy
niloofar_mire · x · 2026-10-21
Startup humans& released Ampersand v0.1, pitched as the first model trained for people to achieve more together, available through its Agents API.
- Different training paradigm: current pipelines with verifiable or preference rewards push models toward autonomy and replacing work; Ampersand is trained as a collaborator that understands individual and collective goals.
- Evaluation: the model is embedded in simulated user groups with unstated, shifting goals over long horizons; the company claims it outperformed Fable and other frontier models in multi-month simulations.
- Emphasizes timing and restraint — when to speak, when to hold back — as equal in importance to task completion.
Related event: Startup humans& launches Ampersand v0.1, a collaboration-first model(7 posts)→
More from Models
- DeepSeek-V4.1-Flash tops Agent and WebDev Arena as #1 open model — arena · 2026-10-21
- Is 1M context plus compaction the endgame for all long-horizon AI tasks? — zmkzmkz · 2026-10-21
- Snorkel AI audit: 55% of TerminalBench 4.0 fixes had reward hacking issues themselves — ravithejads · 2026-10-21
- Ampersand v0.1 Ships via Agent API: Hosted Sessions, OpenAI-Client Compatible — gharik · 2026-10-21
- Ampersand 0.1 Nearly Ties Fable 5.1 on Collaboration Eval, Starting from GLM 5.3 — gharik · 2026-10-21
- Commentary: socially intelligent multi-user agents are the next step to real-world deployment — niloofar_mire · 2026-10-21