This is where we’re at, I guess … the best minds of our generation, all the frantic madness, the resource and ingenuity, and this is the outcome.

The incentives are so strong – the pattern feels intractable. I’m reminded of Diogenes the Cynic.

- Peer-to-peer data storage systems (what I sometimes call "spatial registries") offer interesting new ways of managing geospatial data.

One very interesting use case: self-sovereign policy zone management.
Another: content-addressed climate datasets.

Lots more to say on this ...

- Verifiable geocomputation systems can similarly improve trust and reduce compliance costs.

I'm especially excited about how these two pillars can help us build systems to govern autonomous machines — including safeguarded AI.

www.lesswrong.com/posts/P8Xcbn...
twitter.com/soona/status...

You can dive into what I mean in the paper, but key highlights:

- Proof-of-location systems based on principles of self-sovereignty and verifiability can improve trust in location claims, with many knock on effects

It's worth noting that Vincent Weisser is one of the OGs in DeSci (Molecule ecosystem), with a deep background in the design principles of decentralization.

x.com/vincentweisser

Did you see this article in the Economist last week?

I'm far from an expert on the pros and cons of different training designs, but at a glance decentralized (well, distributed) training offers quality improvements not in spite of, but *because of* inefficiencies.

www.economist.com/science-and-...

... audio processing and more.

www.x25bd.com/posts/retros...

In 2025 I'll be doing research + product innovation using AI and Web3 tools and patterns. More to come ...

... and next generation geospatial data infrastructure.

The last few months of 2024, I reconnected with my roots in data science — I went through an immersive program catching up on the latest developments in AI and machine learning. Vector embeddings, transformers, computer vision, MLOps ...

I've been interested in geospatial for a long time — my first brush with machine learning was in 2009, studying convolutional neural networks in a remote sensing course I took.

Since then I've worked on location-based apps, geospatial analysis pipelines, interactive map visualizations ...

But while I believe Web3 promotes and important set of design principles, I know it's not a panacea.

Before my foray into the space — and in the past year — I've spent a lot of time trying to understand, and to some extent shape, patterns of human behavior using geospatial data technologies.

I'm still working in Web3 — I lead Astral, an open source R+D studio building tools + technologies for the decentralized geospatial web.

I'll share more about what we do in due course, I'd love to get feedback and critique here.

astral.global

About a year ago I left my full time role at Toucan (lots of reasons, mostly personal — I was burnt out). 2024 involved a lot of sense-making and tending to the things that tended to get de-prioritized while I was wrapped up in the "founder" lifestyle. Lots to say here — I'm biased towards openness.

Regarding work, I spent the past 5(ish) years at the frontier, looking at how emerging tech (especially Web3) could help us respond to the climate breakdown more effectively.

I have been deeply involved with @kernel0x.bsky.social, co-founded Toucan Protocol, co-chaired a WEF working group and more.

I grew up in the US and am an American, though I've spent the past 6+ years living in the UK, and just recently took steps towards dual citizenship.

I love the community spirit I experience in the US, and the extraordinary landscapes. In the UK (esp. London), it's the cultural and historical depth.

To say a bit more, I lead with where I have spent my time because my sense of place is core to how I experience the world.

Roots are early experiences exploring mountains + lakes in America. These days, I've expanded — most of my tech work incorporates "geospatial" in one way or another.