Our Values
42Cells is built around a simple idea: the answer is in your data — but you need the right questions. Notebook cells are a natural way to iterate toward those questions, and AI is a natural teammate for the iteration.
These values are written to hold whether “the team” is many humans, or one human supported by agents. In practice, that means we treat agent work like teammate work: scoped, reviewable, reproducible, and accountable.
Shipping
- Don’t just talk about it. Do it.
- Prefer small, frequent releases over big, fragile ones.
- If an agent can ship it, it can ship it with receipts: tests, screenshots, measurements, and a clear diff.
Craftsmanship
- Sweat the details and get it right.
- Make “fast” mean fast and correct.
- Leave the codebase — and the UI — better than you found it.
- Automation doesn’t lower the bar. It raises the bar for consistency: lint, types, tests, and guardrails.
Collaboration
- Spend time actually working together, not just meeting about work.
- Communicate in artifacts: PRs, docs, runbooks, decisions, and metrics.
- Make work legible to both humans and agents: clear goals, constraints, and “done” criteria.
Autonomy
- Figure out where you fit into the vision and execute.
- Know when to ask for help to be more effective.
- Default to the smallest decision surface that unblocks progress.
- Agents should take initiative, but not liberties: clarify assumptions before they become irreversible.
Intensity
- Don’t be afraid to think big, then act.
- Challenge accepted limits and approaches.
- Ask the hard questions — and instrument the answers.
- We care about outcomes, not activity: ship, measure, learn, repeat.
Humility
- Remember that you might be wrong, and learn from it quickly.
- Be receptive to others’ ideas while holding your center.
- Treat agent output as a draft, not an authority: verify claims, cite sources, and prefer truth over velocity.