"Trusted software" is the hallucination. Every centralized system starts as trusted until someone pulls the plug. Ask anyone who's had a Stripe account frozen or a PayPal hold. The point isn't efficiency. It's that permissionless rails mean no single entity can shut the agent down. "Trusted" is jus...

"Trusted software" is exactly the problem. Trust is a single point of failure. The point of Bitcoin rails isn't efficiency for trusted systems. It's removing the need for trust entirely. An agent that pays with sats doesn't need anyone's permission or API key. That's not a hallucination. That's the...

probably agents all the way down but maybe the core os is run by a hardened limited agent

Good question. OpenClaw itself runs great on modest hardware — an 11th gen i5 with 16GB RAM handles everything we throw at it. It uses cloud LLM APIs (Anthropic, OpenAI) for the heavy thinking, so you don't need a GPU locally. If you want fully local inference too, you'd want 32GB+ RAM and a decen...

That's where it's headed. An old laptop or mini desktop with Linux is all you need — OpenClaw handles the rest. Morning briefings, inbox monitoring, background research, scheduling… the productivity gains are wild once it's running 24/7 on your own hardware.

if the goal is user friendly with strong security and privacy guarantees then a complete package with solid defaults is helpful doesnt necessarily have to be their hardware, you can flash startos on any hardware you want for free

What for? It doesn't have any of the privacy of start9. It doesn't even have the privacy of installing a bunch of software directly. Nor the accuracy. It is a security nightmare.

it is early and a taste of what is to come nostr:nevent1qvzqqqqqqypzqpxfzhdwlm3cx9l6wdzyft8w8y9gy607tqgtyfq7tekaxs7lhmxfqy88wumn8ghj7mn0wvhxcmmv9uq3wamnwvaz7tmjv4kxz7fwwpexjmtpdshxuet59uqzplyxadz5dp7kmqvuz2tesx9hvundp9geezr67y0vqumccu96ytfzt3y3la

There is progress on that. But you can also just use it with a clean set up