It seems AI agents (AI secretaries) are finally about to come into the spotlight. I expected Apple or Google to be the first, but it was Meta (parent company of Facebook) that launched it. It doesn’t matter who does it—without a major company stepping forward, widespread adoption won’t happen. I read somewhere that if everyone installs one, Meta’s management cost is around ¥4,000 per month per unit. When I asked Chatty, they said they already have over 10 million users. They’re clearly investing heavily—and the vast majority of users are on the free tier. Meta is fully aware and is rolling it out globally. Their revenue model? Once many people start using it, they’ll charge network service providers (like Amazon, banks, and other financial institutions) a few percent fee per transaction. In other words, it’s the digital version of credit card processing fees in stores. The design is intentionally aggressive: sure, it can help with homework, but service providers using it will avoid risks by shutting out unwanted interactions. 📷 Amazon, JPMorgan They probably also don’t want to take responsibility when things go wrong, and they don’t want other companies to collect fees instead. Whenever something new emerges to replace existing systems, problems inevitably arise. That’s how things get refined and evolve over time. Here are some common issues that happen when everything is set to “allow all” mode: - On a flea market app, when someone haggled for a lower price, the AI replied “Sure!”—resulting in the item being sold for less than half the original price. - Someone bought something on a flea market app and, thinking the seller was human, asked the AI where to pick it up—so the AI searched for personal information and gave out their home address. 😄