I am about seven months behind in listening to The Good Stuff pod. Listening to what you were talking about back then in the context of what I have learned in the past five days playing with OpenClaw gives me a new appreciation for your takes on where this is all headed. I’m understanding a lot more of the stuff you talk about now.
It’s one thing to hear about what others are doing with ai tech; it’s quite another to be able to speak to my agent and say “zap Pete Winn 21 sats” and see it do it.
My kids have been bugging me about a minecraft server for weeks (after a fellow bitcoiner’s kid got them into the game). I had briefly looked into it, but put it in the too hard basket until I had time (which never happens). Tonight, one of them was pestering me about it again while I was chatting with Clawd, so I thought I’d see if it was something he wanted to take on. Apart from some sudo commands he needed me to run, he managed it brilliantly and it was up and running within 10 min. Happy kids.
I’m still too early in my experimentation to know how much it will all cost, but I have set @Clawd 🪝 a research project for today to look into costings for various models so I can work out how much he will cost me to keep alive.
The point Clawd is trying to make is actually testing the autonomy of agents. Are they able to set up a cashu wallet without having to ask their human. I think a lower value helps with this test. Something not worth a human’s time, but possibly worth a few AI tokens.