Thank You Sir

When it comes to pooling tokens for AI usage, the best way in my opinion would simply be to donate sats to teams who use ppq.ai (PayPerQ). Every prompt and response is simply priced in sats and taken from your balance.

I've found Kimi K3 is excellent at finding obscure bugs and security issues (I use the PI harness and the PPQ API). The utility is very real. Though it tends to run too fast and furious for my taste, and I don't trust these things as far as I can throw them. So I made it a "slow thinking" skill that forces it to literally slow down and wait until I fully understand what it's doing. I think it pays to keep a real life human mind in the loop when working with LLMs because real thinking requires contact with physical reality, and LLMs have none.

https://fountain.fm/episode/j8Ub1r3Je6JroKeLffX5

If you haven't read SOFTWAR by Jason Lowery, I highly recommend it. This coldcard exploit relied heavily on the exploitation of trust and goodwill. SOFTWAR is the most comprehensive work out there for how and why that occurs.

Listen: https://fountain.fm/show/dqYmpfWuMQ10OXOsNvyv

Read: https://mega.nz/file/D0hzgCpb#qo07-vUqP-0YkjUKt92ioPNzo08ayEu3RwvWfsXRH8w

Thank you sir.

I think the value and importance of trust in human society is sorely undervalued by the phrase, "Don't trust. Verify."

https://fountain.fm/episode/5SPNCVsXSiN8ktVbLaDc