@antpalkin: AI just took the last corner of trading where a human could still win. 14 of the 20 most profitable wallets on Polymark…

X AI KOLs Timeline News

Summary

Bloomberg data shows 14 of the 20 most profitable Polymarket wallets are AI bots, with over 100,000 humans losing at least $1,000 since January last year. The article argues that AI agents have taken over prediction markets, leaving retail traders at a structural disadvantage.

AI just took the last corner of trading where a human could still win. 14 of the 20 most profitable wallets on Polymarket are bots. Over 100,000 humans there have lost at least $1,000 since January last year. Both numbers counted by Bloomberg. Prediction markets were that corner. No charts, just a question in plain English and whether you know the answer. It fell anyway. Machines already run 75% of US stock trading, and this was the last room where retail still felt even. The bots are not smarter than you. Not by a single percentage point. They are just everywhere. You bet on the two or three questions you were awake for. AI agent plays all of them, reads the resolution rules before it looks at the price, and never has an off week. Wallet 0x8dxd: $313 to about $438,000 in a month. 6,615 predictions, 98% hit rate. The wallet is public, anyone can go count them. Nobody hits 98% by being smart. That is what it looks like when you are simply there first, every single time. None of them is a fund, either. The advantage dies on bets over about $500, too small for a research desk to bother with. It is people running code with AI. Until last year that meant being a developer. Now it means describing what you want in plain English and letting an AI write it. So it stopped being humans against machines. It is you against the guy who spent one evening typing.
Original Article
View Cached Full Text

Cached at: 08/12/26, 04:31 PM

AI just took the last corner of trading where a human could still win.

14 of the 20 most profitable wallets on Polymarket are bots.

Over 100,000 humans there have lost at least $1,000 since January last year. Both numbers counted by Bloomberg.

Prediction markets were that corner. No charts, just a question in plain English and whether you know the answer.

It fell anyway. Machines already run 75% of US stock trading, and this was the last room where retail still felt even.

The bots are not smarter than you. Not by a single percentage point. They are just everywhere.

You bet on the two or three questions you were awake for. AI agent plays all of them, reads the resolution rules before it looks at the price, and never has an off week.

Wallet 0x8dxd: $313 to about $438,000 in a month. 6,615 predictions, 98% hit rate. The wallet is public, anyone can go count them.

Nobody hits 98% by being smart. That is what it looks like when you are simply there first, every single time.

None of them is a fund, either. The advantage dies on bets over about $500, too small for a research desk to bother with. It is people running code with AI.

Until last year that meant being a developer. Now it means describing what you want in plain English and letting an AI write it.

So it stopped being humans against machines. It is you against the guy who spent one evening typing.

Similar Articles

How AI Will Save Prediction Markets (10 minute read)

TLDR AI

The article examines the failure of prediction markets to achieve their utopian vision and argues that AI can transform them by enhancing market creation, analysis, and accuracy, shifting focus from sports and crypto to high-impact questions.