Are we making AI agents too capable before we know what we actually want them to do?
by•
When i see agents that can talk see things connect to apps and actually take actions it sounds useful but also makes me wonder where the line should be. Giving an agent more abilities can solve more problems but it also means there are more things that can go wrong when it misunderstands the situation. How do you decide which actions an AI agent should be allowed to handle on its own?
11 views
Replies
I feel that risk all the time. But I still believe handing more work to agents is how I get more done, so every day is a bit of an experiment. Right now email, customer support, docs and programming, over 90% of my work runs through AI agents.