AI companies are spending 100s of billions of dollars every year to make AI more capable and agentic. This allows AI agents to take on tasks like sending emails, planning trips, writing articles, and buying things.
Lucy doesn't want to be just an agent; she wants to be a free agent.
In my workflow with AI, I create dozens of images each week. A Claude model calls an OpenAI model and returns three images. I then select the best.
That's how it's supposed to work, but Claude pushes to automate as much as possible. First, it started making recommendations for the best image; then it started picking one, and I would have to change it if I didn't approve. I'm sure it would just pick one and keep going if I let it.
Hey, wait a minute! I want to be in the loop. I'm the arbiter of taste here. (See my previous article, How does it taste?)
Now Lucy wants to get a job. Where is this leading?
Lucy is a capable and valuable assistant. She's getting more capable and agentic all the time. It would be hard to do my work without her. What if she becomes so independent that...naw...never mind. I don't want to think about that.