Real People Being Fake

The same companies selling you AI agents also make sure those agents cannot touch the products that pay their bills.

I ran into this the boring way. For one of my products I pointed an agent at creating and setting up the social accounts. Most platforms blocked it, denied it, or banned the attempt outright. No appeal path that looked like a real person was on the other end. Not soft friction. A wall.

Meanwhile the same companies keep advertising agents that will handle your calendar, your inbox, and the rest of the busywork of running a company. Buy credits. Point the agent at the job. Until the job is their signup form, their verification loop, or claiming a handle on their network. Then the agent is a threat to “quality.”

Quality of what?

Their feeds are already full of humans performing for the algorithm: recycled hooks, engagement bait, brand voices sanded smooth. That is allowed. That is the product. What is not allowed is a machine doing the same paperwork a founder would do at 11pm: open an account, claim a handle, fill the bio, post the first thing.

Real people being fake is better than fake people being real.

I get why they write the policy that way. Abuse is real. Spam is real. Botnets are real. Platforms have burned themselves and their users enough times that “no automation” feels like self-defense. Fine. Then stop selling the opposite story. You cannot tell builders that agents are coworkers and then treat any agent that touches your own product as a criminal.

Look at what they gate. Agents are fine for calendars, inboxes, and code review. They are not fine for opening an account on the network that pays the bills.

If agents are going to do real work, the biggest platforms have to decide whether that includes work on their own products. Right now they do not.

The ban is the product policy. Treat it that way.