Agent Operating System

Agents are clearly going to have a big impact on work. They will probably have a big impact on our daily lives.

I can picture lots of ways that I would leverage a reliable agent throughout the day. In fact, I think in the near future, I’ll have at least a couple of agents working on something at all times.

The keyword is reliable.

How will I make sure the agent isn’t making mistakes on my behalf along the way? Or even better, how do I make sure that the mistakes the agent makes are trivial to fix?

I wonder how I will manage all these agents. A kanban board probably will be too simplistic. I’m probably going to need something more sophisticated.

I wonder if I will use multiple agents built by different developers the way I use different chatbots like ChatGPT, Claude, Grok and Perplexity.

If that’s the case, will they be able to interact somehow?

Will I have to manage them all separately across different interfaces?

I sure hope not.

I wonder what the agent operating system will look like.