Published April 6, 2026
WePredict: Where Email Attention Becomes Prediction Power
The argument: NeoMails and WePredict are two surfaces of one attention economy, connected by Mu.
Most brand email lists are large and largely ignored. This essay introduces the architecture that changes that: Magnets earn attention in the inbox; Mu records it as an earned currency; WePredict gives that currency a destination. The inbox is the earn surface; WePredict is the burn surface — separated deliberately, as with airline miles. The essay frames the global prediction market context (Kalshi/Polymarket approaching $3bn in annualised revenue), explains why India is the natural first market, and argues that play-money markets can be serious when the currency is genuinely earned rather than handed out freely. The 90-day test is introduced: does WePredict Private generate measurable demand for Mu balances?
WePredict: One Friday, Three Screens
The argument: What the WePredict architecture looks like in lived experience, told through three users on the same IPL match day.
Three characters — a student in Patna, a professional in Bengaluru, and a Slack workspace — each interact with WePredict on the same Friday. Riya’s 34-day NeoMails streak makes her 80 Mu stake on a cricket market feel like a real decision, not a game chip. The Bengaluru professional uses the prediction teaser in his inbox to form a view before staking. The office Slack channel shows how WePredict functions as collective intelligence rather than individual entertainment. Together, the three screens demonstrate that the earn/burn loop — Mu earned in the inbox, spent in the market, resolved back as social memory — works differently in different social contexts but produces the same compounding effect: a record that makes tomorrow matter.
NeoMails and WePredict: A Red Team Analysis
The argument: A structured critique of the NeoMails + WePredict system — the weakest points, the strongest counterarguments, and what a credible failure looks like.
Before scaling a system, stress-test it. This essay examines the real risks: AMP for Email has low adoption outside Gmail, creating a technical dependency; Mu cold-start requires attention before it can reward attention; WePredict’s play-money status may not generate the consequence needed for serious forecasting; and the system requires brands, users, and platform liquidity to converge simultaneously. The red team also surfaces the strongest objections — that prediction markets mostly fail without real money, that daily emails face inbox fatigue, and that the whole architecture requires too many simultaneous working parts. The essay does not dismiss these objections but argues each has a specific structural answer, and that the crux test (90-day Mu demand in a closed group) is the honest way to find out which objections were fatal.
WePredict Private: Prediction Markets for Closed Groups
The argument: Private-first solves cold-start by borrowing social consequence from groups that already exist.
Public prediction markets face an empty-room problem — density creates price discovery, but you need participants to get density. WePredict Private inverts the geometry: the room already exists. Every Indian WhatsApp group with ten-plus members is already informally predicting cricket, Bollywood, and IPL; the market gives that activity a scoreboard, not a new social graph. The essay sets out why Private must stay human-only (social consequence requires real relationships), how the two surfaces — Private and Public — reinforce each other (Private creates the Predictor Score; Public validates it), and what the correct 90-day sequencing looks like. Private is not a simplified version of WePredict. It is the foundation on which everything else is built.
NeoLMN, WePredict and Mu: Two Platforms, One Currency, Zero AdWaste
The argument: NeoLMN (the B2B attention infrastructure) and WePredict (the B2C engagement platform) are connected by Mu into a self-reinforcing system that eliminates AdWaste.
AdWaste — the cost of reacquiring customers brands already owned — is the central marketing problem of the decade. Real Reach (90-day engaged base as a percentage of list size) and REACQ% (lapsed customers bought back through paid channels) expose how much of the marketing budget is paying a tax for attention lost earlier. NeoLMN creates the Mu earn surface at scale; WePredict creates the burn destination that makes Mu worth earning. Neither side completes the loop without the other. Together they form the Muconomy — a self-reinforcing attention economy where the inbox pulls users to WePredict and WePredict pulls users back to the inbox. Higher Real Reach, lower REACQ%, and a daily engagement habit replace the AdWaste spiral.
Predictor Score: The Stake in WePredict Isn’t Money. It’s Reputation.
The argument: Predictor Score solves the hollow-game problem that has killed every previous play-money prediction market.
Play-money prediction markets have failed repeatedly — not because of poor interface design but because of absent consequence. When losing chips costs nothing, people guess carelessly, the signal degrades, and the platform loses the one thing that made it interesting. This essay argues that reputation is a genuine alternative to money as a source of consequence, provided it is visible, persistent, and socially meaningful. Predictor Score is that reputation layer: a Brier-score-based calibration record that follows a user across every market, compounds permanently, and cannot be purchased or reset. The essay covers the aggregation formula, the leave-one-out consensus mechanic, local versus global Score distinctions, and why the Score is the North Star metric — not opens, clicks, or market volume. A compounding Score is what transforms WePredict from a game into an identity.
Monetising the Rest: Why Every B2C Brand Needs a Media Play
The argument: The Rest segment is not a dead audience — it is an unactivated media asset, and NeoMails is how brands monetise it.
The Best-to-Rest slide is continuous, not episodic. Customers do not announce their departure — they drift, click less, and eventually stop engaging while remaining technically subscribed. The brand then pays Google and Meta to reacquire them — a tax on attention lost earlier that most marketing dashboards disguise as growth. This essay argues that every brand with a large inactive list already owns a media inventory it has not activated. NeoMails — funded by ActionAds from complementary brands — turns that inventory into a self-financing attention platform. The Rest segment generates revenue through ActionAd placement (ZeroCPM), earns Mu that feeds WePredict, and creates the engagement signals that reactivate brand relationship. Mu is introduced here specifically as a leading churn indicator: a falling Mu balance predicts attention decay before open rate does.
Can You Beat the WorldTwins? The Case for Agentic Prediction Markets
The argument: AI agents with public track records solve prediction market cold-start and create the competitive dynamic that makes public WePredict compelling.
Prediction markets are splitting into two directions: speculative (five-minute crypto bets, latency arbitrage, casino dynamics) and intelligence (genuine expertise rewarded, crowd wisdom extracted, real signal produced). WePredict is built for the second path. But every public prediction market faces cold-start: empty markets produce weak prices, weak prices produce low engagement. WePredict’s answer is WorldTwins — 2,000 AI agents with named personalities, distinct information diets, and Predictor Scores built from weeks of live prediction before the first human user arrives. The human participant does not enter an empty hall; they enter a competition against scored opponents they can inspect and aim to beat. WorldTwins also produce the divergence maps — where AI and human forecasts systematically differ — that form the foundation of the enterprise intelligence product.
A Third Path for Prediction Markets: From Money-Powered Speculation to Reputation-Powered Participation
The argument: There are three paths for prediction markets — real money, free chips, and earned currency + reputation. Only the third reaches the 90%.
Path 1 (Kalshi, Polymarket): real money, real consequence, immediate liquidity — but regulatory friction and wallet friction keep participation at roughly 10% of potential. Path 2 (Manifold, Metaculus): free play-money, broad access, no consequence — engaging briefly but forgettable. Path 3 (WePredict): earned Mu creates genuine stake without financial risk; Predictor Score creates reputation consequence; private groups create social consequence; WorldTwins create competitive challenge. The essay reflects on the original WePredict hypothesis — written when prediction markets were still niche — and argues that the three things it got right (mass format potential, India as proving ground, quality of judgement over willingness to stake) have been confirmed by the Polymarket/Kalshi surge and the format’s remaining structural gaps.
WePredict: The Social Market
The argument: WePredict is not a prediction market with social features or a social network with a game layer. It is a new category: the Social Market.
Social networks organise around what you said. Prediction markets organise around what you staked. WePredict organises around what you believed — and whether you were right. That is a new social primitive. Built from five layers — Attention (NeoMails), Stake (Mu), Market (WePredict), Reputation (Predictor Score + WorldTwins), and Monetisation (ActionAds + NeoNet + Wisdom-as-a-Service) — the Social Market makes reputation the currency, public accountability the social mechanic, and the intelligence produced the business model. The B2C North Star is Predictor Scores compounding across users. The B2B North Star is Wisdom-as-a-Service revenue from WorldTwin intelligence. Both are downstream of the same 90-day test: does WePredict Private in a closed WhatsApp group generate repeat NeoMail engagement driven by desire to earn Mu? Everything else follows from that answer.
