Papercompute

Parallel source trees are cheap; parallel runtimes are not. How a disposable, fully isolated platform environment per working tree lets multiple problems - worked by humans or agents - run side by side on one laptop.

How we're building the free and open source building-blocks for the future of AI data

Dreams sound mystical. In practice, they are a way to label agent traces, reflect across sessions, and turn repeated behavior into something the next run can use.

I audited our agent sessions expecting to find wasteful repeat calls. Repeats were only 4%. The real money was in agents re-reading what they already knew, plus a permission-check tax we could finally put a price on.

What we see our AI agents doing may not be exactly what's going on under the hood.

Your team keeps paying for the same solved problem, and because the tickets all close, nobody sees it happening. Repetition is evidence of a missing team capability, and standup is the detection layer that's been announcing it all along.

London's bridges are failing under load nobody modeled. Haussmann rebuilt Paris for traffic that hadn't arrived yet. Agent infrastructure faces the same choice: skills, shared team sessions, and reflections are how you build for the traffic that is coming

Before companies use AI cost overruns to justify cutting people, they should terminate the agent workflows that consume money without producing reliable outcomes.

One session, $100.66, 91% Fable 5 - for confirming rollout steps. Multiply that across 200 engineers and you have a budget conversation nobody is ready for. The paper console makes the numbers visible.

Evidence gets a skill written. Evaluation keeps it worth keeping. A skill that isn't measured can silently make your agent worse, and, because skills execute, a wrong one does more damage than a wrong doc ever could. Here's why every skill needs an eval,