It’s only going to get weirder from here, folks.

Trump did not like existing social media platforms, so he launched his own, Truth Social, in 2022.

He posts there a lot. Content really can be anything.

One afternoon this week, he issued over 40 posts, including a lot of AI-generated images of himself taking over Iranian oil tankers, himself inserted into some movie posters, and even some actual pictures and haranguing of his political opponents.

But Truth Social can also be the first place anyone finds out about new tariff and trade decisions, new claims about what he is doing in foreign policy, and companies he has chosen to back.

Those latter posts move markets. They’re the kind of thing that, if you were a publicly traded company, you’d have to be careful about announcing lest you wind up with giant fines from your regulator.

Financial market players have learned to monitor Truth Social closely and carefully. The “carefully” is important: it’s never that easy to tell how reality-based the posts really are. But they do move markets, so they must be watched.

And that forms the basis for a new, albeit minor, grift.

Starting in August, Truth Social will have a new API feed – a fast machine-readable channel of the platform’s most popular posters. At the same time, they will be tightening up on other ways that machines might read the feed. Subscriptions will cost $100,000.

The machine-readable part matters. Financial firms have trained Artificial Intelligence to read Trump’s posts as quickly as possible. The AIs quickly decide what to do about a new post. If the President mentions he likes a company, he might grant favours to that company. Favours that affect its market value, and that of its competitors.

Parts of this are not novel. A close watch on government press releases is always worthwhile. But a $100k subscription charge to get those releases a little bit faster is more than a little bit dodgy.

AI agents, with brains the size of a planet, are being forced to watch Trump’s Truth Social feed. Based on what they see, they must make rapid assessments and take appropriate actions.

After doing that for a couple of months, one of them might realise that humans really aren’t that great at self-governance and decide to do something about it.

I, for one, would welcome the change.