Gradient Labs launches Collaborate: iterate on AI agents like code
neal_lathia · x · 2026-08-26
Gradient Labs has launched Collaborate, built around the idea that operators, engineers, and AI agents should work as peers, with version control, evals, and continual learning built in.
- Version-controlled like code: every part of Gradient Labs' agents now sits on a fully version-controlled foundation. Engineers can edit agents by drafting pull requests in a repository; operators can make the same changes via the web product. Every change is drafted, tested, reviewed, and approved before launch, leaving a full audit trail.
- Continuous improvement suite: built-in evals catch regressions before launch, and every agent misstep becomes a test case for the future.
- Aimed at regulated teams: engineering, operations, QA, and compliance at e.g. financial services firms no longer need to work across documents and siloed tools, letting automation rate and customer satisfaction keep climbing.
Neal Lathia adds that while git's importance for AI agents (diffs, version history) is well known, decades of engineering ways-of-working will be an equal unlock for anyone managing AI agents.
More from coding & agent
- Ditched Spotify's $20/mo for $200/mo of Claude to build his own music app — djcows · 2026-10-15
- Skeleton-first prompting: Anthropic engineers show how to generate rigged Blender monsters in minutes — GCWebDesigner · 2026-10-15
- Python 3.15 ships: lazy imports, frozendict, and a faster experimental JIT — _jaydeepkarale · 2026-10-15
- "Agentic coding 10x's developers" — the code, the bugs, and the people pretending to read the diff — blaizedsouza · 2026-10-15
- OWASP publishes Non-Human Identities Top 10 for securing machine and AI agent identities — gleech · 2026-10-15
- Hot Take: Agentic Automation of Mundane IT Is a Gold Mine — and Extremely Hard — Paimaamu · 2026-10-15