Protocols are the lever. Keep context open, durable, and user-centric, and we get competitive AI that serves people first.

Many commercial digital products now function like public utilities. Long data histories and well-worn workflows create real switching costs — with scale comes a duty to stay reliable and user-centric. Strong incentives to protect control and capture value create challenges.

Join millions who have switched to Grok.

One failure mode is API gatekeeping: brittle terms + policy whiplash that rug-pull competitors. This is where the durability of these architectural decisions is important: if possible, lock it open with technical enforcement; otherwise, regulate.

If context is portable (with consent), big platforms can still aggregate data and accrue value. But that same context, which I argue belongs to the user — can flow to an open network of products competing to serve the user. Innovation without platform lock-in.

A core proposal by the authors: secure, open APIs with user-delegated access. This would put the context that makes AI useful under the user’s control — and make it portable.

I read it through the lens of the Web3 design principles I outlined lsat week: open, durable, opt in. In that sense, it’s a blueprint for a “Web3 approach” to AI architecture. (This might not jive with some definitions of "Web3", which is ok — there's certainly room for debate).

This AI Frontiers piece on open protocols from Isabel Moure, @timoreilly and @IlanStrauss at @aif_media is a clear articulation of why open, durable, opt-in design matters. They provide a counterweight to the centralizing gravity of unregulated digital platforms.

RT @maceagon: really enjoyed diving into @AstralProtocol with @johnx25bd -- his reflection here about "moving thoughtfully" via @Kernel0x r….

The internet can be an alienating place — Kernel proves that it can also be a place of connection and community. More than just "crypto", Kernel is a place for anyone thinking deeply about how to build technology that helps us live well. If you have questions, I'm available.

I've struggled with the term "Web3". So confusing, so emotionally loaded . I've slowly come to view Web3 not as a set of products, protocols or networks, but as a set of design principles:. - Open.- Durable.- Opt-in. Read 👇:.

@TaylorOshan 10/. If any of this resonates — if you're working with location data and need it to mean something — I’d love to hear from you. We’re still shaping the future of this space, and looking for thoughtful collaborators across domains. 🧭.

@TaylorOshan 9/. Next up is a focused period of customer discovery and validation. We want to learn:. – Where does this approach stand up?.– Where does it fall down?.– Who actually needs this?.– What would make it easy to adopt?.

@TaylorOshan 8/. We’re excited to apply this in domains where credible location evidence really matters: insurance, logistics, identity, content licensing, public safety. When location is tied to money, compliance, or reputation, weak evidence creates real risk.

@TaylorOshan 7/. We’re developing a system that takes raw location evidence — GPS, networks, sensors — and turns it into cryptographically signed location proofs. Each proof is a verifiable digital record of where and when something happened, scored based on the signals it draws from.

@TaylorOshan 6/. One simple insight: stacking multiple independent signals produces stronger location proofs. It’s like multi-factor authentication, but for place. GPS alone can be spoofed. So can Wi-Fi or IP. But when multiple signals line up, it gets much harder to fake.

@TaylorOshan 5/. We’ve been building a framework to make sense of these systems — evaluating them by precision, cost, privacy, trust model, threat resistance, and more. The goal is to make their trade-offs clear, and help people identify the right approach for their needs.

@TaylorOshan 4/. There are many different types of PoL systems. They rely on different sources: GPS, Bluetooth, Wi-Fi, device sensors, secure elements, social witnesses, legal contracts and more. Each has strengths and weaknesses. No single approach works everywhere.

3/. For the past year I’ve been working with Prof. @TaylorOshan at UMD, studying proof-of-location (PoL) systems. At their core, PoL systems produce evidence that a location claim is true — evidence that would require technical manipulation, collusion, or fraud to forge.

2/. It’s easy to measure location (e.g. using GPS) + report it. But making credible location claims — especially when there’s an incentive to lie — is harder. And it’s everywhere: ridesharing, delivery, dating, streaming, fieldwork, voting. Location is a basic digital primitive.

1/. I’ve spent 7 years researching where geospatial data and decentralized systems meet. One challenge keeps coming up:.How do you prove where something happened?.Online, verifiably, and in a way others can trust?. 🧵:.