I spent the past few weeks building deep context about myself for my agent, with the goal of making that context portable across any harness/model.
My expectation with Astra was:
Give it goals, context, and permissions → let it decide what needs to be done → proactively work toward those goals without me defining every schedule/task.
So far, I haven’t been able to make that work.
Schedules still feel like “wake up and execute this predefined thing” rather than “wake up, understand my goals, decide what matters, and act.”
Maybe my agent architecture is wrong. Maybe I’m expecting too much from current models.
For people actually building autonomous agents: what am I missing?
Or is “give an AI a goal and let it run” still mostly AGI hype?
@kellyroman@adcock_brett A lot cheaper then hiring a company to clean your house, take care of your yard and fix your car (if we still own cars in the future)!
“Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us.”