An OpenClaw early user on painful setup and fading daily use
The author says deploying OpenClaw took many hours and that they now use it less and less, keeping only a few cases like shared memory, daily updates and event watches, and predicts the new crop of assistants will face the same retention problem.
Hot take? I barely use openclaw anymore… When it came out I deployed it (many hours, very painful) But it was so cool! Over time I found I used it less and less. A few core use cases remained: a central shared brain/memory for all my agents (not really an agent job TBH), a daily new updater, event watches. Will this happen to folks using the latest crop of agents (instinct, muse, dots, etc) I’m wondering what has changed with this latest crop of agents, or whether people will find the same pattern that I’ve seen. Long term daily active use is the key. Honestly with @alexandr_wang and the team at Muse working on this I bet they’ll have some of the best discipline on monitoring and iterating through this problem. Facebook and instagram were built in this same way, painstaking discipline on what drives engagement and iterating like crazy. I predict The future if personal agents will look very different than it does today.