Someone used an AI probability model to make $2.2 million in two months on Polymarket—this story is definitely eye-catching. Fully automated trading by a robot sounds like a groundbreaking strategy.

But honestly, I’ve always been skeptical of such news. Over the years, I’ve heard too many claims about AI trading robots, and in the end, they’re either obvious scams or the technology simply doesn’t hold up. Automated trading sounds very sci-fi, but in practice, various issues keep cropping up—network latency, market slippage, model overfitting… just one or two factors can wipe out the profits.

As a prediction market, Polymarket itself has limited liquidity and trading depth, making it more susceptible to large orders impacting the market. If such arbitrage opportunities truly exist, the question is: how long can this advantage last? Once more people find out, the opportunity disappears. Also, what exactly is the logic behind the robot’s strategy? Has the market structure changed? None of these details have been disclosed.

Maybe some people really are making money in prediction markets through clever algorithms, but given the current environment full of scams, it’s best not to get carried away until verified by real data.