Anthony, an ex intelligence officer turned entrepreneur, joins Sunny Ray to trace an unusual path from dotcom era tech jobs to CIA style fieldwork in Iraq, Asia and Africa, then into the National Geospatial Intelligence Agency as its first CTO, Bridgewater's core management team, and Cerberus Capital as a managing director. He explains how those experiences converged into VCO, an AI and data science platform that forecasts political, economic and geopolitical events with probability scores, not just narratives. Anthony details how VCO predicted US and Israeli strikes on Iran and the closing of the Strait of Hormuz days or weeks before prediction markets like Polymarket and Kalshi caught up. He describes why large language models alone cannot forecast reliably, why he left a comfortable private equity job to build VCO with a former Citadel and Google colleague, and how the company pivoted from a personal hedge fund idea into a broader forecasting tool meant to help everyday people make better life decisions, from careers to college choices, not just traders and governments.
An ex CIA case officer and Bridgewater alum built an AI that called the Iran strikes before the prediction markets did.
What are you building at VCO?
We wanted to build something that would forecast the future better than people. We built an AI and data science system that gives the probability of a political, economic or geopolitical event and explains the reasoning behind it. What's special is we've built a suite of tools around it to help people actually make decisions, whether that's trading a stock or, in my old world, a military or national security call.
You predicted the Iran strikes. How did that happen?
We launched the system in January, on our site and on the Bloomberg terminal. We were tracking whether the US or Israel would strike Iran and whether the Strait of Hormuz would close, and we called all of it days or weeks before Polymarket and Kalshi did. You can see the charts on vico.io or on our X and Substack. Seeing is believing.
Are you using the same LLM tech everyone else is, or something different?
It's both. You have to use LLMs, they're so powerful, but LLMs alone aren't great at forecasting because they're not deterministic enough and can get caught up in their sources. We combined LLMs with mathematics, data science, and modeling and simulation. Our team has straight up AI guys, mod sim guys and data science guys working together.
You started in tech around 1999. What did that first ride teach you?
I was at a company called LinkShare doing affiliate marketing, back when e-commerce was just starting. I literally read a book called Introduction to E-commerce to get the job. What stuck with me is that technology can help people understand the world and make better decisions, and that idea has been a throughline through intelligence work, Bridgewater, and now VCO.
What pulled you from startups into becoming an intelligence officer?
The proximate cause was 9/11, I wanted to serve and protect the nation like a lot of guys my generation. But honestly I'd always been fascinated by that world and wanted to be a spy at some level, even though it doesn't feel like a real career option. It turns out you can just go do it if you're interested.
What did the field in Iraq teach you about decisionmaking that no classroom could?
As a case officer, my job was to find people with useful information, recruit them, and handle them as they went out and collected it, then I'd write it up for generals and presidents. People think it's glamorous but 80% of the time is spent writing reports. You do it with weapons and armored vehicles for protection, and with teams who are deeply dedicated to the mission.
What did radical transparency at Bridgewater actually feel like from the inside?
It's a social experiment and I wanted to see if it could work. Ray gave a name to radical transparency and made it a conscious part of the workday, not just an aspiration. The biggest lesson was idea meritocracy, training yourself to actually listen when an intern says your idea is dumb instead of shutting them down, which is much harder than it sounds.
What made you decide to leave Cerberus in 2025 and start VCO?
AI is a once in a lifetime technology, like the app revolution after the iPhone. In 20 years we'll be explaining to our kids what life was like before AI and they won't understand it. Once my cofounder and I landed on the idea, we quit our jobs and went all in.
Building something daring? Sunny talks to founders like this every day. Fifteen minutes to see if your story belongs on the stage.
Claim your pre-interview