I'm enjoying the concept. Tight browser integration/control. Controlling agents by commenting on Artifacts (Task List, Implementation Plan).
But the user experience for basic things has just been difficult. Unresponsive Agent Manager window. Agents hitting file permission errors because I don't have a proper "workspace". Agents getting blocked waiting for me to confirm a command line's "y/n".
A great step forward for those who are already locked into the language! But IMO not necessarily encouragement for newcomers to start using the language for production software.
Generally agree, using LLMs to expand on your work is usually just boilerplate or garbage.
What about using LLMs to refine or sharpen your existing work? Similar to a Rubber ducky? If you're intentional about maintaining and understanding the theory behind the work, I've found it a useful tool.
But the user experience for basic things has just been difficult. Unresponsive Agent Manager window. Agents hitting file permission errors because I don't have a proper "workspace". Agents getting blocked waiting for me to confirm a command line's "y/n".