Joseph Kim · AI Engineer · 18:59.
Close the gap between model ability and real work
- Kim’s three product pillars are financial data, a review-friendly interface and firm-approved output templates.
- OpenAI licensed and indexed data itself after finding problems with existing connectors; models still reach it through MCP tools.
Make outputs inspectable
- Source viewers expose the passages and tables behind an answer.
- Charts, spreadsheet models and presentations fit into existing review workflows.
- Kim contrasts tool instructions with post-training that builds familiarity with a dataset.
Measure usefulness after review
- A task crosses the usefulness threshold when checking the result costs less than doing it yourself.
- Persistent agents and expert-level financial forecasts are Kim’s future bets.
- Caveat: this is OpenAI’s product account and roadmap, not an independent test of investment performance.
You need to be able to check where a number came from and understand the context around it.