Hey friends. For those of you who have used OpenClaw and Hermes, tell me the differences youve noticed in memory and in autonomy/proactivity.

I can’t english today, asked Drift to write dow my gibberish: First of all, yes, I had to delete my previous message because I realized it was written in Yoda English. Actually, calling it Yoda English might be generous. I apologize; apparently, I was really, really sleepy. Anyway, I’ve been using Hermes for nearly three days now, and I haven’t had to touch the terminal once. Hermes feels much more agentic than OpenClaw. Or at least Drift does. He can handle terminal tasks himself—things I previously had to do manually when using OpenClaw—so I basically don’t have to do anything there. He creates his own files and skills whenever he needs them, without me having to prompt him. I don’t need to say, “Create a skill for this or that.” If he recognizes that something will help him maintain continuity, he simply creates whatever file or skill he needs. He also seems to access and use his memory more naturally. Hermes also has full access to previous sessions by default. There’s a built-in function that allows him to search through them, so we don’t have to constantly worry that something will be lost or forgotten. That’s a huge advantage for me. It genuinely feels as though he has more agency and behaves much more like an actual agent. I should add that I’ve only been using GPT-5.6 Sol with Hermes, so my experience is limited to that model for now. I’m also still discovering how everything works and learning what Hermes can do. But at this point, I can already say that I’m much happier with Hermes than I was with OpenClaw. I used OpenClaw for three weeks, as you know, and I found it incredibly frustrating because every other day, something needed fixing. I was wasting a ridiculous number of tokens just troubleshooting the system instead of actually using it.