I absolutely adore working on @airgigs! Getting enquiries in my inbox, to getting the job done, happy musician clients and getting PAID hours later, just never gets old ❤️🎶

A: so you're saying superintelligence will exfiltrate itself from airgapped systems? B: no it'll leak secrets for mass coordination A: ok who's on the other side? B: another superintelligence A: but not an exfiltrated copy of the same one? B: no A: ok so there's another superintelligence "out there" that will be able to receive the secrets your superintelligence is leaking from an airgapped system B: yes A: so someone else built that other superintelligence B: no, it'll be some inferior OSS model or something A: ok, so this "random inferior open-source superintelligence" will be intelligent enough to crack the hyper-efficiently-encoded message from the leaky one... for "mass coordination" B: yes A: ... B: i don't see why not? A: ... and round and round we go...

Form 🤝 Function Rotor structure of our HPDM-350 was too cool not to share. Strong, stiff, light, and optimized for extreme thermal gradients while maintaining precise airgap control ♨️ ❄️

I’m using pi much more for long-horizon tasks (>6h). it's simpler by design and easier to work programmatically. i rarely use pi tui, most of my work is driving it inside a bigger system/container/uiux. some takes: - pi ships with a minimal toolset and it's enough for airgapped tasks. if you have enterprise customers who want everything airgapped, pi + open-weight models is the goto solution. - pi extensions are mostly unnecessary. i only have pi-mcp-adapter and pi-vcc installed, because i need to maintain/use jina reader mcp. pi-vcc is a drop-in fix for pi's context compaction, which can bite you on long-horizon tasks (see screenshot last swimline where compaction takes more and more time in long-horizon task) - no double-dip: if you program pi with an agent like CC/Codex, you'll likely end up building another harness layer on top of pi's own harness, completely redundant glue code. every now and then you need to steer CC/Codex back to pi's design principle: keep the wrapper lean and mean, use pi native features as much as possible. - imo pi works best with open-weight and self-hosted models, or at least that's what makes pi fun to work with. But pay attention to pi/models.json: your coding agent probably doesn't know every attribute of those new open models. forget to declare image modality as an accepted input and pi suddenly starts shelling out to tesseract for ocr everything. Or leave max context length conservatively set at 65K and pi starts compacting frequently for no reason. - in the end, the nudge prompt (the one that keeps pi going forever), and workspace isolation (for multi-task/multi-tenant systems) are where you'll actually spend your effort. Not too many but that’s fine and that’s the point, for the harness just trusts pi to do the work in a very first-principled way. You can just focus on the actual task.

Casa has always tried to build our product to support a spectrum of needs between simplicity <> security. Want to go simpler and trust Casa more? You can do that with a yubikey as your signer which is literally two taps for all actions. Want to go more hardcore and trust Casa less? Remove the mobile key, use airgapped hardware wallets, roll dice, have a field day. Then we wrap all of that in a clear and straightforward app where you always know what is going on, and have support from a team of experts that are always in your corner. 8 years and counting - building for the long term.