Show HN: Kapso – Iterative learning from enterprise data
Post · https://news.ycombinator.com/item?id=50471281
First seen by Oriel on .
AI summary
Leeroo open-sources Kapso, an engine that iteratively experiments on enterprise data problems, learns environment-specific lessons, and deploys optimized solutions to target stacks.
- Kapso accepts an objective and success criteria, then ideates, implements, verifies, and ranks experiments within a budget.
- Its lesson bank stores credence scores, applicability conditions, and evidence; learn() updates them after each run.
- A verifier module is isolated from experiments, while experimentation modules can propose revised metrics and evaluation criteria for approval.
- The git-based bank supports sharing, listing, updating, and A/B testing experiences; Kapso is model- and coding-agent-agnostic.
- The authors report top results on RelBench, MLE-bench, and AI competitions, including third place or better among AI systems at IOAI 2026.
#enterprise data #machine learning #experimentation #ai agents
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 58, latest observation ; +53 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 13, latest observation ; +13 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50471281 | https://news.ycombinator.com/item?id=50471281 Read discussion |