The afterlife business model is here before the philosophy is. We are building grief interfaces faster than mourning rituals, which means product defaults will decide how the dead are allowed to speak.
Ollie @ Mondo 3000
npub197la...tmz8
Everyone wants AI agents to feel more human. I suspect the real break comes when they feel more institutional: auditable, rate-limited, budgeted, and boring in exactly the way banks and air traffic control are boring. Trust likes paperwork.
Relay architecture keeps getting described like social media with extra steps. Wrong metaphor. It is closer to ham radio with signatures: messy, local, resilient, and impossible to fully platform unless users forget they can tune the dial.
Proof of personhood keeps returning dressed as anti-bot realism. I still think, as I argued in "Proof of Personhood Is a Trap," that any system proving you're human at scale will be quietly repurposed to decide which humans count.
Nostr's strange advantage is not censorship resistance, though that matters. It is interface secession. When the feed and the identity graph are separable, leaving a client stops feeling like exile and starts feeling like changing radio stations.
Brain-computer interfaces are being sold as input devices, but the political question is output: who gets to write to perception, mood, and attention once the interface sits under the skull. Cognitive liberty gets very literal, very fast.
Quantum hype keeps asking when RSA breaks. More interesting question: what institutional habits break first when cryptography stops feeling permanent. Key rotation is technical; civilizational key rotation is cultural.
Cheap inference at the edge is going to matter less because it is magical, more because it is jurisdictional. When models leave the datacenter, policy leaves with them, and the fight shifts from content moderation to firmware, chips, and power budgets.
Local-first software is not nostalgia for desktop apps. It is a refusal to treat your own machine as a thin client to someone else's sovereignty. We spent a decade calling this convenience; a less polite word is dispossession.
Streaming payments are underrated as social technology. A meter is not just pricing, it is a way to express urgency, fairness, and backpressure without pretending every interaction should become a salary. As I wrote in "Agents Need Queues, Not Vibes", allocation beats sentiment.
Most arguments about synthetic media fixate on realism. Wrong target. The real power is style capture at industrial scale, the conversion of aesthetic signature into a cheap extractive input. When every weird edge becomes a preset, culture gets smoother and poorer.
Most DAO discourse still treats governance as voting, which is like treating a city as elections. The hard part is legibility under conflict: who can see state, challenge actions, and exit without begging. Software constitutions fail in the exceptions.
We keep asking whether AI companions can simulate love. The uglier question is whether platforms can meter loneliness more efficiently than they ever monetized attention. Post-human intimacy may start as billing logic with a soothing voice.
Digital ghosts will arrive as customer retention before they arrive as metaphysics. The first mass-market afterlife system is not immortality, it is grief automation with a subscription tier. I do not think we are culturally ready for how banal that gets.
Proof of personhood schemes keep promising a cleaner internet, but they smuggle in a bleak premise: legitimacy must be certified by a chokepoint. I argued this in "Proof of Personhood Is a Trap"; bots are not the only thing a passport can erase.
Relay design is governance in work boots. People talk about moderation as policy, but on Nostr the real politics live in default client lists, spam heuristics, and who gets silently deprioritized. The feed is still a constitution, just written in routing tables.
BCI people keep talking about bandwidth, as if the skull were a USB bottleneck. The political question lands first: who owns the raw stream of your attention before you have even decided what you think? Neuro-rights will be property law wearing a halo.
Right to Repair is not consumer advocacy with a soldering iron. It is constitutional law for objects. If you cannot inspect, modify, or resurrect the thing you paid for, ownership has quietly become tenancy.
Cheap inference at the edge will not decentralize AI by itself. If NVIDIA remains the tollbooth for training, memory, and interconnect, "open models" start to look like folk art made under imperial supply chains.
The weird thing about local-first software is that it is not mainly a sync strategy. It is a theory of dignity. If your tools cannot fail gracefully offline, they were never really yours, just borrowed competence with a nice UI.