February 2026: OpenClaw
Ooh shiny!
Easy install. Check.
Hmmm something isn’t working right.
Some time later, I got it fixed.
Ooh an update! Installed. OpenClaw broken.
Some time later, I got it fixed.
Hearing good stuff about Hermes. Friend raving about how it self-learns. I’m like “Yeah, but OpenClaw has a huge community and skills can give it self-learning.” I counseled myself to be patient.
Ooh an update! Installed. OpenClaw broken again.
Some time later, I got it fixed.
Pattern? Yep, turned out to be a pattern.
I delved deeper. Oh… OpenClaw is built on a buggy, fragile foundation.
Hearing an increasing amount of pro-Hermes talk. Hmmm. A friend who operates on a deeper technical level than I do tells me more about how fragile and insecure the OpenClaw foundation is. I’m convinced enough to give Hermes Agent a try.
Oh and on the positive side: OpenClaw is open source. So is Hermes Agent.
April 2026: Hermes Agent
It rocks! Night and day. Faster and more solid. And built with Python, yay!
It’s been almost a month now and I’m having a blast with Hermes! It really does learn. Also, frequent updates bring excellent improvements that *don’t break the system*! I’m thinking (a) Hermes Agent is built on a foundation with far more integrity than OpenClaw; and (b) the primary devs rock!
Now back to giving my 3 (soon to be 5 and then 20 and then…) Hermes agents more daily tasks to do.
What about Claude?
If you want to buy into a closed-source walled garden that sucks up your data, overcharges and under-delivers, and locks you in to using only their LLM models, sign up with misAnthropic. I sure won’t.
Watch the animated musical story / prediction of where things could be going and how we get there. Trigger warning: It starts out dark.
Recent Comments