We are entering an era of information-accruing assets. We're seeing different approaches to this: @supertake uses AI agents to translate natural-language investment takes into managed stock portfolios. @peopledotfun / @cleopetrafun were exploring thesis indexes of prediction markets. @PerpsPadfun backed the indexes by perps. But I believe the real challenge in this category is building the intelligence capable of translating an abstract thesis into an accurately constructed portfolio of market positions. A single thesis can imply dozens of outcomes across completely different markets. - Which outcomes actually support the thesis? - Which markets offer the best exposure to those outcomes? - How should capital be allocated across them? - And how should those positions change as new information emerges? These aren't simple questions that can be solved by matching keywords or asking an LLM to pick a few positions. The entire mechanism is only as good as the intelligence deciding where the capital goes. A poorly constructed portfolio can lose money even when the underlying thesis is correct. You need to understand the information behind a thesis well enough to construct a portfolio that accurately rexpress it through the right positions, at the right prices, with the right exposure. This is the wedge of my experiment. We built an OSINT intelligence engine focused on filtering through the noise, connecting fragmented information, identifying relevant signals, and translating broad speculation into evidence-backed positions. It consistently translates complex, evolving information into positions that extract more value from the information those assets represent. It's a race to turn information into alpha.