待翻译：Credentialism is coming for AI agents

AI 服务暂时不可用，以下为来源正文，待恢复后补全翻译。

This morning I was reading Scott Belsky’s point on favored agents. As he puts it: “What if the moat for these agent companies is to achieve ‘favored agent’ status with direct preferred access to services like Uber, Airbnb, Amazon, Amex, and the long-tail of companies that otherwise guard against bots?” And I think it’s actually part of a much bigger question: who gets to act on our behalf online? The way I’m thinking about this is that, yes, there will probably be some bigger-stakes regulation that shapes how this plays out. But either way, we will definitely continue to see AI agents being blocked from accessing certain marketplaces and services. This might be surprising for people who were expecting AI agents to bring more personal agency and freedom, but credentials for agents will likely be a big thing. The same way the university you attended (or didn’t) and the last company you worked for matter when you’re trying to open new doors. For example, you might have some banks start to allow OpenAI agents or GrokBot to access their APIs. Or maybe those aren’t the ones. Maybe they’ll only allow Microsoft’s Copilot, because that might drive customer adoption or come with some existing enterprise relationship. It’s going to be interesting to see whether this concentrates even more power. Do we end up in a world where only big tech companies are able to get this kind of access? Or does this become a new race where individuals can apply for some kind of cybersecurity certification, so that a personal bot they’ve built themselves can be trusted and allowed to access these services, and banks and other companies can then choose to let it in? There is definitely potential for a whole new layer of gatekeepers.