Now I am excited by MCP and would be all in except security.
Security is a huge issue.
Forget AI and imagine a system where you call APIs and you get both data and JS. And that JS executes at global scope with full access to other APIs. And so do all the other MCP servers. Furthermore the MCP server may go to arbitrary Web pages and download JS. And that JS e.g. from a strangers Github issue or Web search gets executes with full API privileges.
This isn't something MCP can fix. It is built into the dice rolling nature of LLMs. Turning predictions into privileged executions. And those dice can be loaded by any MCP server.
Or imagine surfing the Web using a 2001 browser with no protections against cross domain scripting. Then having a page where you choose what init scripts to run and then it cascades from there. You are logged into your bank at the time!This is what worries me. It's not USBC. It's sort of USBC but where you are ordering all your peripherals from Amazon, Ali express and Temu and the house is made of tinder.