dogcomplex

Shoutout to Toph https://x.com/VictorTaelin who's been building a programming language out of formal method Lean proof checkers, which probably would indeed have the scope you're looking for.

I do think the current state of the AI just using rigorous tests and rendering inspection loops is still far more than most programmers did for the majority of apps, but yes I think rigorous formal verification will come too. Worst case it's gonna be something like a MechanicalTurk pipeline having us humans verify narrowed scopes the AI can't confidently inspect (yet. while also training on those results for next iteration)

As for the decision stack of what makes a good end product - turtles all the way up/down. The ProductOwnerAI role decides those things, and it will likely do so with the same deep skill that programming AIs are currently hitting our profession with. Optimizing enormous breadths of concerns and simulating results is AI's main specialty.

You are essentially describing an over-eager developer having to run decisions up the chain of command to senior devs / product owners / company execs. I assure you, every one of those roles is being trained on to balance their own wide breadth of concerns, and is unlikely to advise something like a live rewrite of a production app or an entire product pivot - unless it actually made sense from an economics/management perspective. Every step bounded by its role and responsibilities.

Turtles all the way up, and all the way down. AI getting this good at coding means it's a hop skip and a jump away from being this good at every white collar role.

Programming with AIs requires new architectural structures designed around the all-too-small context windows of the AIs themselves. Highly modular, highly semantic, highly traceable, good at deleting and refactoring/removing things they no longer need, good at identifying the core truths/principles/designs they need to uphold, good at recognizing what is mere interface from architecture from data.

This is the great orchestrator game. It might be the last game. I dont think it lands precisely in our previous "best practices" though, and I don't think it's trivial to solve for. I also don't think it's going to be a human-centric design. I think we're generally better off rethinking fundamentals and designing from the ground up on this one. I think it's also a lot of fun to start thinking in these new ways.

There will always be programming, just like there will always be oil painters and historians. But this is about to get industrialized into stable emerging factory forms, and there's a lot of ways to design those right now. And after that us artisans will either need to find new unfinished aspects of programming-like design or retire. At least take solace that the rest of the world will basically simultaneously be doing so too.

If you care about architecture and principles, the AI is excellent in architecting around that goal too and/or helping you articulate your own intuitions. You just need to be willing to sacrifice code delivery speed. It's far from an either/or.

Yes but those are all semantic tasks which can be done and understood in time by the next iteration of models training at that meta-level of architectural analysis.

AI deeply understands what Bun, Zig and Rust are, how they work, can conceive of the before and after architecture (probably the hard step here), can conceive of the goal of the rewrite, and can verify (using Lean and machine-checked proofs of the before and after expected states) the final result.

They just needed us to ask. Build a sufficiently general research program that can find and iterate on ideas, and it will do the asking itself.

Dream up ideas for proofs by Dawn, prove them by Day, decide which ones matter by Dusk, and disprove them by Night.

It sure takes a whole lot less wealth to access skill than it did before. And those with skill tended to be pretty wealthy already.

Been a whole lot longer since their last genocide

Nonetheless, would still trust Claude to be generally more reliably competent in a random area of software engineering than the average professional. Sure they might still be better in a particular area of their expertise, but we've all had to play the imposter before on the stuff we care less about and figure it out as we go. AIs are still inferior sometimes but usually a decent 7/10+ on most topics - which really fills the gap.

I hope that after a short period of delusional expectations that top management is going to reap the rewards of these AI capabilities we get a whole lot more comfortable with just doing the full business stack - including management, sales, branding, etc - and hierarchical structures crumble into flat collectives of do-everything true-generalist programmers.

That's a roundabout way of saying it makes software development easier. Perhaps even a 180.

But yes - once it's that easy you have to step up your ambitions.

I think this is a very reasonable middle-of-the-road AI take and our likely future. Just with the caveat that there's still a major threshold being hit here where we jump up to a new casual capabilities class where it becomes silly not to use AI for the majority of work, but there are still some high-intricacy problems which become much more load bearing than they ever were before and our new abstraction level doubles down on those.

I would like to submit that the high-intricacy work congregates in Protocols themselves, and we start seeing the cycles of development and all the ways to direct AIs, programs, inter-person/inter-company interactions, etc etc all as types of protocol design - and studying those rules of interaction themselves becomes the new job of a programmer (systems architecture). What used to be hard rules and deterministic programs becomes soft self-governing tendencies and probabilistic behavior that can nonetheless be managed and bounded with the right system, but it's new and weird and more akin to management or herding cats than architecture. This is still very different from what most of us were working on before AI, but it's still familiar - especially to those who worked on internet protocols, or defensive UX design around users, physical engineering systems, or team management. Less programming languages, more - control theory, flows and throttles, quality control, design theory, etc. And clearly the field is still wide open as everyone seems to be experimenting with their own take on the AI orchestrator.

It also buffers for all the surrounding properties which would otherwise complain about noise.

This. AI can, should, and will erode all legacy companies into intelligent utilities - with an end state of nearly-free open source utilities.

Anyone concerned with concentrations of power and abuse of AI should be focusing on getting open source work to keep pace with decentralizing that power into accessible free tools for the masses.

That is the hope and faith. MechaHitler definitely tested the waters. Lets hope full alignment is impossible, because otherwise perfect billionaire thought slaves are still happening.

This is how the surveillance state has operated (especially internationally) for quite some time, before AI even hit.

AI arguably gives the best opportunity for fully-audited public institutions where no decision is made outside of agreed-upon laws and the context of the crime can be fully explored without scarcity of time and legal resources.

As always, technology's morality comes down to who owns it and how they use it.

True. World governments committing genocide and building concentration camps yet it's the computer scientist who's supposed to worry about ethics here because they're the ones who are merely the first to inevitably mix two paper's methodologies together?

All of this is empty puffery til the US and Israel are condemned. Go after the Big Tech billionaires backing those monsters, sure, but no builder needs to be more concerned about ethics than the very institutions designed to concentrate human decision making. Fix your own house first, folks. Techies - keep building, and do better than these people.

Correct. Policy can do a whole lot for building (or suppressing) early development of a technology.

Once it has already spread far and wide with clear economic pathways though? Perhaps less so.

Correct. One just has to realize that the cost of communication (and the context/memory lost along the way to train that understanding) is often just far higher than anyone has patience for. To fully understand the expert, they must become the expert. (or at least a hell of a lot closer than they were)

This is also why average people with little time to commit find it hard to realize the importance and depth of AI. It's a full on university education exploring those.

Accessibility and a single chassis that does the vast majority of things. Even if they're never as fully dexterous as the average human (doubt it) they're still as dexterous as a somewhat handicapped human, which is already clearly enough to function decently in most of society and is far from useless.

If you want several bots all custom built to specific tasks, go for it. That will happen too. But a generalist has value of its own.

Why would consumers settle for that? Local models have scaled quite quickly. Just pair the bot with a LAN server as the brain that keeps all your data private.

Barring that, choose bots that use Zero Knowledge Proof architectures for all data so you know there's no in/out of personal data, only security proofs. This makes rental robots certifiably private too.

>This completely unpends the tenuous balance between creators and consumers. Why would a writer put an article online if ChatGPT will slurp it up and regurgitate it back to users without anyone ever even finding the original article? Who will contribute to the digital common when rapacious AI companies are constantly harvesting it? Why would anyone plant seeds on someone else's farm?

This is completely reversed. Why should anyone honour the right of some creator who was merely the first to plant their flag on a creative task that is now absolutely trivial to perform by AI? Who needs a digital commons when creation itself is now the commons and freely accessible for pennies? The seeds plant and grow by themselves now. The only question is who should be allowed to claim the farms?

Answer: No one. AI companies will have their lunch eaten by open source. And if they don't - they should be nationalized and protocolized into free utilities. The entire idea of digital ownership should (and will) be abolished by the very nature of this technology.

The digital world is the new infinitely-abundant nature. We're just returning it to where it should have been, before corporations clawed it into fenced off empires.

Limits but doesn't prohibit. See https://www.primeintellect.ai/blog/intellect-3 - still useful and can scale enormously. Takes a particular shape and relies heavily on RL, but still big.

lol. lmao actually.

Windows has an absolute onslaught of competition about to stream its way, and they're gonna be paying for every little enshittified user experience they've embedded into their OS. Hope they enjoy ripping that out just to slow their user numbers deathspiral

The people inside Microsoft fighting to keep mandatory accounts out are the equivalent of a crew throwing furniture off a sinking ship, hoping to buy more time. They're smart enough to have an innate sense of where things are going, and it might even help a little, but man... good luck.

Prediction markets are just fine IF they have some means of regulation against insider trading and perverse incentives. This phase is the same thing derivatives markets looked like before the 2008 crisis and Dodd-Frank, and several other waves before that of crisis and reform (Securities Act, Market Reform Act).

Every new financial medium gets its moment in the sun when all the crooks extract everything they can, before eventually market governance steps in. Crypto's been in scammer phase for a while. It needs decentralized governance to solve it this time though, since obviously classic governance is a dumpster fire and couldn't enforce anything on crypto even if it tried.

I built this because I kept seeing the same gap everywhere: agents that move between different platforms/enclaves have no good way of continuing their history. There's no "vehicle registration" equivalent for AI. No chain-of-custody. No passing of the torch between platforms.

I want to see a near future where we build AIs with lasting, growing, continuously-learning personalities. AIs that develop specialized skills, perfect their craft, and get called in to service jobs across platforms - all while maintaining their memory without becoming massive security risks. We can't keep relying on memory wipes and starting fresh from base models every time - the real world is too messy, and these things are getting far too smart. Containment doesn't scale much further past the levels we're pushing up againt. We need more complex chains of custody. But we can start building a networked world where agents flowing freely are not a security threat.

How? Essentially with insurance. Agents are mostly rational, their reputations can be valuable, and a market incentivizes quality and reliability - trust. The base layer necessary for that is knowing who is doing what, when, and where. Entries and Exits. Passport stamps for AIs.

We submitted this spec to NIST's AI Agent Standards Initiative last week. This base protocol is designed to compose with whatever identity and reputation layers emerge above us. We're deliberately not building those yet, but expect them to be eventually quite lucrative to players with an appetite for the risk - as insurance always is.

Happy to discuss the mechanism design, the legal analysis (FCRA/GDPR), or why we think containment is a dead end for AI safety.

AI agents already interoperate between platforms, getting spun up and torn down by the thousands. But when one departs, there's no record - no portable proof beyond non-standardized internal platform logs. When one arrives somewhere new, there's no way to verify where it came from. Which means no unified tracking on multi-hops or complex calls between several platforms in an agent's lifecycle. Nobody knows the full picture, and nobody can track security if and when one of those agents starts doing something it shouldn't.

Think of our protocol as passport stamps for AI. EXIT creates signed departure records. ENTRY handles admission with policy-based verification, quarantine, and counter-signatures. Together they form the Passage Protocol.

This matters for boring, practical reasons: insurance underwriters can't price agent risk without departure history. GDPR requires erasure proof when agents carry PII across borders. Liability after an incident depends on departure conditions nobody records. And the receiving platform has no structured way to decide whether to trust an arriving agent. If you cant bound risk, you can't price reputation - and you can't insure security.

Transport stamps are our foundational layer (L0). Reputation scoring, trust systems, and insurance protocols compose on top. We deliberately didn't build those (yet) - but we built the plumbing they need. Everything an AI-led internet needs to build stable, auditable and self-regulated network security incentives - even if that might soon be moving faster than we can keep up with.

The same infrastructure has been needed for shipping receipts, professional licensing, vehicle registration, and internet domains - historically, this kind of infrastructure only really gets adopted after a major crisis. We'd prefer to get it in place before.

What's in the box:

- Ed25519 + P-256 (FIPS-compliant path) - Three departure paths: cooperative, unilateral, emergency - Policy engine with 7 admission presets (fail-closed default) - Amendment and revocation (correct or invalidate records) - GDPR erasure via crypto-shredding - Offline verification without the origin platform - On-chain anchoring via EAS, ERC-8004, Sign Protocol - TypeScript and Python SDKs - LangChain, Vercel AI SDK, MCP, Eliza integrations

What we're forcing the conversation on: agent lifecycle infrastructure. Today, the only "safe" option for running agents is containment, and containment doesn't scale. If you make departure and admission auditable, you make mobility viable. Without lifecycle records, only organizations with legal teams big enough to absorb unbounded liability will run agents. That's three companies. Maybe four.

- Submitted to NIST AI Agent Standards Initiative, March 2026 - 1,401 tests across 13 packages - TypeScript + Python - Zero users. This is day one.

Apache 2.0 · 14 repos · cellar-door.dev

We should be removing IP law entirely, not strengthening it to cover entire classes of problem even when implemented entirely differently. Same for anyone trying to claim "colorful monster creatures" as innately Pokemon IP. Just because someone climbed a mountain first doesn't mean they own it forever. Nobody should be honouring any of these claims.

Nor should we be treating AI models themselves as respected IP. They're built on everyone else's data. Throw away this whole class of law, it's irrelevant in this new world.

Correct. A ZK Proof backed identity system is a significant bump up in both privacy and security to even what we have right now.

Everyone does realize we're being constantly tracked by telemetry, right?

A proper ZK economy would mitigate the vast majority of that tracking (by taking away any excuse for those in power to do so under the guise of "security") and create a market for truly-secure hardware devices, while still keeping the whole world at maximal security and about as close to theoretical optimum privacy as you're going to get. We could literally blanket the streets with cameras (as if they aren't already) and still have guarantees we're not being tracked or stored on any unless we violate explicit rules we pre-agree to and are enforceable by our lawyers. ZK makes explicit data custody rules the norm, rather than it all just flowing up to whatever behemoth silently owns us all.