Essays & research
One engine, every argument
Most AI hands you an answer and hopes you don't ask how it knows. The design philosophy behind a universal debate engine — one reasoning substrate that argues both sides of any claim, in any field, scores the evidence, and shows its work.
Read the essayAI Valuations & the Upstream Intelligence Thesis
Nokia had working touchscreen smartphones in 2007 and lost $190B anyway — not on execution, but on a single unexamined assumption. A structured argument for why the market is measuring AI value at the wrong layer, and what that means for how to build, invest, and compete.
Read the essayThe science behind multi-agent AI
Why a debate between role-differentiated agents — Thesis, Antithesis, and a neutral Synthesis — produces more calibrated, better-cited answers than a single model. A plain-language tour of the peer-reviewed literature, with every claim attributed to its source.
Read the pieceHow we benchmark the engine
A score is only as good as it is honest. We grade Superthesis on the dimensions that actually matter — accuracy, calibration, and citation quality — against gold data like AVeriTeC, and publish the method, the live numbers, and the caveats.
See the numbersMarkets & quant
The same discipline — argue both sides, score the evidence, show the work — pointed at financial markets: a factor study, a strategy design with a live backtest, and a field map of where edges actually live.
The TRIAD Book
A diversified crypto book across 14 majors — five price signals, a crowd-positioning fade, and a perp-spot basis trade, blended and adversarially audited. Held-out Sharpe +2.9. The full tearsheet.
Read the tearsheetThe Trend + Carry Book
A market-neutral crypto book — a change-in-delivery trend signal plus a beta-neutral funding-carry sleeve across 46 perpetuals. Positive in five of six years at ≈zero correlation to Bitcoin. The full backtest tearsheet.
Read the tearsheetBetting Against Beta
The market's safest stocks quietly out-earn its most exciting ones. The landmark low-volatility study, the contrarian edge it implies, and five low-beta names to hold into Q4 2026.
Read the noteThe Monthly Compounder
Designing a pooled, monthly-rebalanced multi-asset strategy for one goal — the most profitable months possible. 68% green over 2021–2025, with a leverage dial and an actual daily backtest.
Read the designThe Futures Edge Tree
A field key to every futures & perpetuals edge — 32 setup families around one root, each adversarially scored as a game, and charted by where the crowd has already arrived.
Read the mapMarket Watch
We run live prediction markets through the engine — argue both sides, score the evidence — and surface where our calibrated read diverges most from the crowd, with the reasoning and sources behind each. See the full divergence digest →
Will there be exactly 0 confirmed VEI 4 or higher volcanic eruptions worldwide in 2026?
The crowd gives this ~69%; after arguing both sides, superthesis lands far lower — about 12%.
Read the breakdownClarity Act signed into law in 2026?
The crowd prices this at ~42%; after arguing both sides, superthesis lands lower — about 27% — driven by a documented collapse in the bipartisan talks that the bill's own advocates needed to clear a 60-vote Senate threshold.
Read the breakdownWill the Democratic Party control the House after the 2026 Midterm elections?
The crowd gives this ~81%; after arguing both sides, superthesis lands lower — about 71% — discounting for redistricting-adjusted seat arithmetic and the possibility of a wave that clears historical lower bounds but still falls short of a gerrymandered map.
Read the breakdownTesla and SpaceX merger officially announced by December 31?
The crowd puts a Tesla-SpaceX merger announcement at roughly 40% by year-end; after stress-testing both sides, superthesis lands lower — about 30% — because two legal preconditions remain unresolved and one has been refuted.
Read the breakdownFDA approves Retatrutide this year?
The crowd prices retatrutide's 2026 FDA approval at 12%; after arguing both sides, superthesis lands at 6% — the regulatory clock alone makes a same-year approval arithmetically unreachable under any standard pathway.
Read the breakdownFed rate hike in 2026?
The crowd gives this ~59%; after arguing both sides, superthesis lands at 55% — a near-coin-flip where the origin of inflation, supply shock versus demand, is the central pivot.
Read the breakdownOpenAI IPO before 2027?
The crowd prices this at ~53%; after arguing both sides, superthesis lands slightly higher — about 57% — with the analytical interval running from 49% to 65%.
Read the breakdownAnthropic IPO before 2027?
The crowd gives this ~71%; after arguing both sides, superthesis lands close — at 74% — with both reads anchored by Anthropic's June 2026 confidential S-1 filing and a market window that is, so far, wide open.
Read the breakdownBillionaire one-time wealth tax passes in California election 2026?
The crowd gives this ~21%; after arguing both sides, superthesis lands marginally lower — about 19%.
Read the breakdownWill the US confirm that aliens exist before 2027?
The crowd prices this at ~10%; after arguing both sides, superthesis lands at ~11% — two independent methods, one skeptical conclusion.
Read the breakdown