Hello everyone, After 20 months of researching and writing, I’m excited to present to you my book: “Newsmakers: Artificial Intelligence and the Future of Journalism” by @ColumbiaUP. Coming to book stores this Fall. Available now for pre-order!🤖+📰=🚀 buff.ly/2Jw2m6F

For 200 years, news+information companies got paid to find what's true and to deliver it. AI makes delivery free. That's half the business, gone. What's left is the valuable half: being first to know something. Original Intelligence. Everyone's racing to build the pipes. The water is what gets scarce. appliedxl.com/research/origi…

This hit #1 in three Amazon bookstore categories in the first week. Didn’t expect that. I’m giving away 10 signed copies. Tag someone in comments who always needs to know things first. If I pick your name, you both win a signed copy. I’ll choose 10 this Sunday. What if you could see the early signals before they become news? That’s not a hypothetical. That’s what this book is about.

It costs $211,900 to produce a news investigation. It costs $0.02 for an AI to extract its value. 10 million to one. That is the First-to-Know Tax. People debating whether AI can "do journalism" are missing the point: It doesn't need to. It just needs to wait 3 minutes for a journalist to do it first. Within minutes of publication, AI has crawled, indexed, summarized, and served the essential value, often before the reader even considers clicking through. News aggregation extracted attention. AI extracts comprehension. The user can now understand the story without ever visiting the source. The parasite needs the host. The host is losing money. But the host hasn't realized it still has the ultimate leverage. A handful of AI companies have started paying publishers. It won't matter. A licensing fee is a toll on a road that's being rerouted. The moment an article is published, its value can be extracted by anyone. No deal changes that. For 200 years, the business model of journalism was temporal advantage: knowing first. AI collapsed that advantage to a window of minutes. The monetization window after publication is 3 minutes. The monetization window before publication, the detection phase, is days to weeks. One window is collapsing. The other is where the entire value of the information economy is migrating. The only information AI can't extract is information that doesn't exist yet. The ratio for a daily news story is 158,000 : 1. What is it for a research report? A legal brief? An equity analysis?

vibe coding took over software: trust the AI, ship the output, don't look too closely. It either works or it doesn't. And it almost always works. The same thing is coming for newsrooms and nobody wants to say it. Call it vibe reporting. Point AI at thousands of documents, let it surface what matters, let it draft the first version. The journalist becomes the editor of a machine. The only question is whether the person editing the machine knows enough to catch what it gets wrong. This isn't a prediction. It's an inevitability. But there's a difference. Code has a compiler.. if it runs, it runs. Journalism doesn't have that. There's no compiler for truth.. yet.

There's a hidden tax in the information economy. Call it the First-to-Know Tax. You do the work of discovering something, the investigation, the FOIA requests, the source cultivation, the data analysis... Then an AI summarizes it in 2 seconds and the value migrates to whoever has the best distribution layer. The organizations that discover information are subsidizing the ones that repackage it. The math does not hold up. Whoever figures out how to monetize what they know before anyone else can summarize it wins.

Last week, CNN's Chief Data Analyst Harry Enten stood before a data wall and pointed to a number that felt like science fiction: the odds of the United States purchasing Greenland. On Kalshi, the prediction market, that probability had tripled in days. Traders weren't reacting to a policy briefing. They were recalculating the future based on the audacious Venezuela operation. For them, the poll was context. The market was the signal. We have entered an era where the first draft of history is being written as prophecy. From election outcomes to Supreme Court rulings, Americans are no longer just watching the news, they're betting on it. But prediction markets are only one element of the age of anticipation. Most surprises leave traces. Market crashes, clinical trial failures, geopolitical ruptures, they look sudden in hindsight, but the signals were there: buried in filings, logistics data, registry updates. Visible to anyone watching the right streams. This unprecedented visibility isn't equally distributed. It creates what might be called temporal inequality. The hedge fund has the algorithm to detect the tremor; the pharmaceutical executive sees the enrollment slowdown months before the trial collapses. The rest of us have the anxiety of watching the clock. We built tools to see the future. The question is whether they give us anything to do with what we see. Welcome to the age of anticipation.

In the predawn hours of Jan. 3, a digital footprint in Arlington, Va., hinted at something big. An X account tracking the "Pizza Meter", late-night delivery surges near the Pentagon, flagged a 90-minute spike. Hours later, Nicolás Maduro was in U.S. custody. Nobody traded on the pizza signal. But someone, somewhere, knew first. They always do. The pattern is Gulf War–era. The technology isn't. What used to be a Domino's manager counting boxes is now an algorithm parsing traffic data and mobile pings. And while the Pizza Meter is public, the signals that move money rarely are. We've turned the "Not-Yet" into a trillion-dollar industry. Predictive systems now flag which turbine will fail, which drug trial is drifting toward collapse. Geopolitics is messier, but even there, logistics data and shipping manifests offer early warnings to those who can afford the feed. The first draft of history has become the first draft of prophecy. But the future isn't equally visible. The hedge fund pays for satellite imagery and proprietary data streams. The pharma giant spots enrollment problems in its own trial months before investors do. The retail investor? The patient in a failing trial? They find out when everyone else does, on CNBC, after the stock has cratered. This is temporal inequality: not just who knows, but who can act. Seeing a signal is worthless if you can't reposition a portfolio, redirect R&D, or get into a different trial. Foresight without agency isn't power. It's a front-row seat to your own obsolescence. Welcome to the Age of Anticipation.

Breaking news is old news. By the time a headline hits your screen, the outcome is largely locked in. Markets have moved. Crises have formed. Decisions have already been made, often by those who didn’t wait. Information moves at machine speed, faster than humans can comprehend. The advantage no longer belongs to those who react best to the news, but to those who recognize what’s taking shape before it becomes news. My upcoming book, The Science of First: Catching Early Signals Before They Become News, explores that narrow window, the moment before events harden into headlines. Co-authored with @ScottMAustinAustin and @ByNikitaRoy examines how weak signals buried in data precede market shifts, systemic risk, and global events. The book offers journalists, investors, and anyone who depends on timing a blueprint for seeing shifts early, acting before the crowd, and building a durable information edge. This isn’t about predicting the future. It’s about noticing what’s already underway while there’s still time to act. Breaking news is outdated. Welcome to the era of Pre-news. Book drops soon. If you’d like early excerpts, updates, or invites to live discussions as we launch, sign up here: appliedxl.com/book

The next era of journalism begins when reporters stop waiting for companies, governments, and research institutions to speak, and start tracking the small operational shifts they leave behind, the filings and schedule changes that reveal what’s happening long before anyone announces it. @NiemanLab Predictions for Journalism 2026 niemanlab.org/2025/12/pre-ne…

We used to operate in a world where decisions followed analysis. Now, analysis follows reaction. That’s the Reaction Economy, where the first interpretation drives the outcome. After the train derailment in East Palestine in Ohio a few years ago, the initial official frame of a "controlled burn" was beaten by a faster, more viral frame of a "toxic cover-up." No amount of subsequent EPA data could reset the public’s reaction. The first frame wins. In the 2024 proxy fight against Disney, the company’s immediate reaction was to frame activist Nelson Peltz as a destructive amateur. That story set the terms of the fight, and all his later analysis was filtered through that initial, skeptical lens. Disney won because its reaction became the story. Reaction sets the benchmark. In the Reaction Economy, speed isn’t optional. It's the baseline. The only advantage left is how clearly you can see, and respond, before everyone else catches up.

We no longer just disagree on facts. We disagree on what’s worth noticing. Two people in the same boardroom can receive entirely different feeds and walk away with opposite interpretations of the same event. We've seen this play out in the OpenAI leadership crisis: one side saw a corporate coup to oust the CEO, the other saw a board trying to prevent AI from going rogue. Or take GLP-1 data: some biotech investors saw it as routine, others as a metabolic revolution. This divergence isn’t abstract, it disrupts coordination. Investment committees, editorial teams, and strategy groups now operate on desynchronized baselines. This isn’t just polarization. It’s desynchronization. Teams can’t align if they don’t see the same version of reality in real time. Decision velocity increasingly depends on synchronized signal flows, not just access to data.

Reality is soft. You can poke through it. For much of the 20th century, institutions like newsrooms, universities, and regulators acted as filters. They validated, slowed, and debated information, so we could agree on what was real. That model is breaking. Now, a single image or phrase from an anonymous account can move markets before institutions respond. We’ve seen this dynamic before, like when an AI-generated image of an explosion near the Pentagon briefly tanked U.S. markets. The image, though fake, moved faster than verification systems. Reality hasn’t vanished, it’s just become porous. Signals now slip through and move systems before verification catches up. In high-velocity domains, momentum tends to begin where perception forms before verification.