Innovation always comes with setbacks, but what happened here is a real problem for the AI x crypto space and a big reason people have written it off. One of the agents in Virtual got hacked for half a million dollars because a developer still held the key that controlled the agent’s smart wallet. The whole thing ended up being a human in the loop system dressed up as an AI agent. This is why many have stopped paying attention to AI x crypto. It is not serious. Today, most AI agents fall into two buckets. - Entertainment agents that generate attention in hopes of people buying a token. - Service agents that hold user funds and execute DeFi actions on their behalf and use the fee to buy its token Across both buckets you have the same problem. Little transparency. Few proof the agent is actually doing the work. Low onchain guarantees that remove the need to trust the operator. Too many past examples where it ends up being humans behind the curtain. Most importantly, if the developers walk away, the agent, the token, everything just dies. This isn’t a dunk. It’s the reason we focus so heavily on building hard guarantees first on EigenCloud. If AI agents in crypto are going to be more than lipstick on a fad, they need verifiable execution and intelligence. Without that, the entire category stays unserious. I’m confident Virtual is thinking in this direction and we are down to jam on ideas. And this setback doesn’t change the long-term arc. AI agents are going to reshape everything and crypto has a real role to play. aGDP