Pick the AI names you hold. Get one number on the same 0–100 scale as our AGI-2027 Thesis Tracker — so you can read your basket directly against the thesis it is priced on.
The score only changes when a verdict actually changes — so this is a low-frequency email, not a newsletter treadmill. Free.
Subscribe free →Each prediction carries the same weight the Thesis Tracker gives it: supportive = 1, open or pending = 0.5, refuted = 0. A ticker's score is the weighted average of the predictions its AI valuation rides on; a basket is the average of its tickers. Inverse positions score 1 − weight, because they pay off when the prediction fails.
That is deliberately the same arithmetic as the Tracker, which is why 62.5 is a meaningful line to compare against rather than an arbitrary benchmark. Verdicts come from /data.json (CC BY 4.0) and this page is regenerated whenever one changes.
The part to argue with: which predictions each ticker is mapped to is an editorial judgement, not a filing. It is an argument about what a company's AI valuation depends on. Holdings are not — those come straight from public 13F filings and nothing is inferred beyond them. There are no prices here and no performance claims: this measures what a basket is exposed to, not what it is worth.
This tool is free and stays free. The recurring version is a different product, and it is not built yet:
There is no waitlist form and no pre-order. The briefing below is where it would be announced, and subscribers get asked what it should cost before anything is priced.
Subscribe free →Free to embed on any site. It carries one link back here.
No. There are no prices, no targets, no buy or sell calls, and no performance claims anywhere in this tool. It measures one narrow thing: how much of a basket's story depends on AGI predictions that are currently holding up, using verdicts published openly at /data.json. What you do with that is your decision, and you should talk to someone licensed before acting on any of it.
Where do the holdings come from?Public SEC 13F filings for Q1 2026, as of March 31, 2026, already audited into this site's investor profiles. Only Berkshire Hathaway and ARK Invest have line-by-line filings here; the other six investors tracked on /invest have qualitative stances only, so they are left out of the holdings table rather than guessed at.
Why does my infrastructure basket score higher than the thesis?Because the two predictions those companies ride on — compute scaling and AI capex — are both currently graded supportive, and capex is graded as exceeded. The capability predictions further out, including AGI by 2027 itself, are still unresolved and score 0.5. A high score means your basket is levered to the parts already working, not that it is safe.
What happens to the score if a verdict changes?Every basket moves. The page is regenerated from /data.json, so a verdict flip propagates to this tool, the Thesis Tracker and the homepage in the same commit. The one prediction already graded Wrong — that open source would fade — is why inverse positions like Meta and Alibaba score the way they do.
Can I save or share my basket?Yes, and it needs no account. Your basket is written into the page URL as you pick — copy the link and it reopens exactly as you left it, for you or for anyone you send it to. There is no cookie, no login and no email involved. Bookmarking that link is also what makes the planned alert useful: when a verdict changes, the same URL will show you a different score.
The Invest section — how 8 legends are positioned · AGI-2027 Thesis Tracker — the score this is read against · Is the AI capex a bubble? · The fund that bet the thesis, and blew up · All free tools