Is AI good enough to run an entire investment process? Let's find out.
I wanted to take 20+ years of investment experience and see how far AI could leverage and compound it — building the investment infrastructure I always wished existed, from the investor's viewpoint.
A team of geniuses working together to get the research and analysis we need.
A lot of people are trying to use AI to make investment decisions autonomously. This isn't that. It's about making the investor better — sharper research, richer context, less time on grunt work. The judgment stays with me.
↳ No product. No pitch. Just the build — in public. Take what's useful for your own process.
↳ The only ask: share what you discover on LinkedIn. Tag me — I read everything.
The team
seven agents, each with a roleHow the team answers a question
the full pipelineThe whole point: you ask, the team goes to work.
One question. A grounded, data-backed view in minutes — not days. Bull case, bear case, honest answer. Built to inform a real allocation decision, not just summarise the news.
The full stack
where it's all headingThe build
filling in as I goNotes
observations as the build unfoldsShort observations posted as I build — what the data is showing, where the model is surprising me, what I'm thinking about next. Follow along and add your own take in the comments.
Just read this after it was flagged by FT Alphaville — worth your time. The vanishing footnote story is exactly the kind of thing the wiki is built for: information that's technically public, buried deep enough in filings that nobody finds it. Until you have a system that does. Read it here →
Follow the build
Posts on what I'm building, what's working, and what I'm learning. No pitch. No product. Just the process.