78733e…b266fb

Software building Software is only the obvious first step. It will grow exponentially and surpass human intelligence. Although I am not in favor, I'm afraid it will be inevitable. Then it will be machines building machines, because the intelligence to do so will no longer be an obstacle. None of what those can replace is safe. If anything is safe, it is activities that require a soul, empathy and humor -- human to human relations -- but I would even add a question mark to that, because man will feel utterly miserable in second place and will want to join the beast. I wish I were wrong.

Lot's of cope in this thread.

> That key is used to encrypt data throughout the process so we can't see your data. Unpopular opinion: It has been bothering me for a while. Statements like that are slightly misleading. Obviously, the data has to be decrypted before it is passed to the model. So transport may be encrypted, and also you may encrypt context data at rest, but there is no such thing as a large language model that can process encrypted data. There is nothing that technically keeps @Maple AI from tapping into the data. Ultimately, it's a trust model, not a technical solution. Prove me wrong.

I don't think so. The next generation does not understand why they need to signup somewhere, why they would need an Email, or why some detached entity would have the right to exclude them from content. Hang in there. It only takes one killer app. Only a matter of time, imho.

Of those, proof of humanness is the only category that leaves little room for interpretation.

Bitcoin's success did not need professional product designers, either. Nostr needs product designers as much as Bitcoin does. Neither is a product. Both are protocols each of which allow a network to develop organically, and there is no better proof of genuine superiority than organic growth. Nostr apps do need good product design, just like hardware wallets do, but that's a totally different issue. You have built products yourself @franzap. Honest question to you: What were the obstacles for better product design?

You can tell openclaw was written by a multi-millionaire. Token usage was definitely not the top priority.

Experimenting with running openclaw in sandbox with local model. It forces you to be efficient with context, and more thoughtful with workflows, because the model is weaker. It's been slower and needed some help occasionally, but so far there was nothing it could not do: research on web, news aggregation, task management, programming tools for itself, posting to nostr, etc. Curious what will happen once I crank it up.

That's pretty close to the distribution of fiat money.

Big lesson here: > Why do you think your work, Open Claw, won? > It’s hard to compete against someone who’s just there to have fun. Peter Steinberger on Lex Friedman 00:22:19