Living theses: a probability that moves with the evidence
Investment theses written from cited evidence, with pillars, kill facts that fire on new data, and a probability that updates as each new fact arrives.
Request access- 72%Google Cloud maintains or grows its 80% AI product adoption rate and 90% Fortune 100 Gemini usage
- 61%Daily token consumption continues to scale from 105T levels with accelerating paid enterprise workloads
- 55%Low per-token pricing combined with volume and smart routing still delivers sufficient monetization to reach the target run-rate
Google Cloud AI product adoption stalls or declines
Written from evidence only
Ask a question, or let rejo's idea engine propose one when strong evidence piles up on a topic. The thesis is drafted only from numbered claims it was given; every mechanism step and risk cites them, and an independent audit checks the citations and the reasoning.
Pillars and kill facts
Each thesis splits into the pillars it depends on and the kill facts that would prove it wrong, with metrics and thresholds where they exist. When a new filing, print or claim crosses a kill fact, the thesis is flagged immediately.
A probability you can explain
Every new piece of evidence is judged against each pillar; the update itself is fixed arithmetic (log-odds with capped steps, weighted by source quality), so every move in the probability has a line in the ledger that says why.
A track record
Theses are resolved against what happened and scored, so you can see how calibrated the desk (and you) have been.