How Much Can You Safely Delegate to AI Agents?
How much can you delegate to agents?
Trusting AI agents depends on the task, not just model quality. I propose a four-level framework based on how easy work is to check and undo. From simple assistance to self-driving modes, this guide helps you decide when to delegate safely without breaking things. By breaking tasks down and using deterministic checks, we can maximize autonomy while maintaining control.
Trusting your agents just because the models got smarter is like skipping your seatbelt because you got a nicer car.
- SOLAR_FIELDS
This is a very long form way of saying “understand the work you are dispatching, and if you don’t, ask questions until you do”
- jolaflow
Interesting. When it comes to auditing, I think there is also another layer to be verified - the intent. We have tooling for static analysis and automated tests, but drift in intent or vision is hard to detect.
After leaving agents unsupervised for longer periods of time, the biggest questions tend to be "how did it arrive at this conclusion?" and "did we drift from the initial vision?". Typically these questions are hard to answer as you only review the final state of the workflow.
I recently built an issue tracker that does offer time-traveling as a core feature. Somewhat unexpectedly this turned out to provide a much sought for overview allowing for a better workflow audit, as detailed here:
https://dev.to/ljtn/vision-drift-addressing-the-next-problem...
Would be interested to hear your thoughts. I guess vision drift would classify as something that is hard to reverse, because you often don't realize it has happened until much later.
- ChicagoDave
I work on my MacBook in direct terminal sessions or in a container with no safeguards.
I save the container for well-planned units of work that can easily be repeated and validated on a branch.
I use direct sessions when I expect and the plan dictates that Claude might go off target and I need to hold the steering wheel.
Fable has increased my use of the container option.