What should we work on?
Ten minutes on where I came from, what I keep coming back to, and what I think is worth building with the tools we have right now. Eighteen slides. Given at the Internet Identity Workshop retreat at Solstice Grove, 7 November 2026.
@image: /anselm/talks/iiw-2026/img/wask.webp @credit: Mount Rundle (Waskahigan Watchi), Alberta. My photograph. @sub: IIW Retreat · Solstice Grove · November 7, 2026
What should we work on?
Anselm Hook. Ten minutes on where I came from, what I keep coming back to, and what I think is worth building with the tools we have right now.
Given at the Internet Identity Workshop retreat at Solstice Grove.
@image: /anselm/talks/iiw-2026/img/louise.webp @credit: Lake Louise, Alberta, 2006. @section: Where I came from @sub: 1976 to 2010
Grew up programming
- Started programming fairly young at my father's computer store in 1976.
- Video games were my real training: the consumer end, the "app" end. Real products, real users. Dragon's Lair. Lord of the Rings. Very, very pretty 3D for millions of people, with artists and designers, under pressure.
- Changed course at Planetwork, 2000. Xerox PARC: digital twins of people. Mozilla Research: studied money. Early crypto.
@layout: site @image: /anselm/talks/iiw-2026/img/glacier.webp @media: /anselm/talks/iiw-2026/img/five.webp @section: Intrinsics, and the anchor @sub: 1976 to now
What should we work on?
What can we really do today, with today's tools? What are the millennium software challenges? The general goal underneath everything: help people see better.
Asked without attachment to past work, the essay lays out five challenges. Identity. Modeling and simulating the world. Molecular engineering. AI alignment. Agency. The second one is mine. The rest are context.
anselm.substack.com/p/what-should-we-work-on
@layout: site @image: /anselm/talks/iiw-2026/img/camping.webp @media: /anselm/talks/iiw-2026/img/figure.webp @section: The thing we are trying to see @sub: 2026
Land use, water, food, energy, migration. The couplings are real and they are not small. People can hold maybe three of these at once in their heads.
Software can hold all of them.
A systems map of climate feedback loops, from the essay. Everything connects to everything, and most of the loops run the wrong way.
@layout: site @media: /anselm/talks/iiw-2026/img/hyperobjects-cover.jpg @section: The thing we are trying to see @sub: 2015 to now
Hyperobjects
Timothy Morton's word for things so widely distributed in time and space that no person ever sees the whole of them. Climate is the canonical case. The cover is an iceberg, which is the right picture: the part you can see is not the part that matters.
Modeling is how you get to see the underwater part.
@image: /anselm/talks/iiw-2026/img/fitz1.webp @sub: 2026
Goals today
- Simulate the world. I think nobody has actually done this. Let's do it.
- Ship a very simple, very focused appliance: help people model near-term outcomes of land use law and policy.
- Solve the core piece only. One appliance, one piece of software. Do not boil the ocean.
- Modeling and simulation are just another consensus tool. Like discussion forums. Like voting. Just an app. Don't get caught up in it.
@layout: site @image: /anselm/talks/iiw-2026/img/laguna.webp @media: /anselm/talks/iiw-2026/img/simulate.webp @section: 2015 @sub: 2015
simulate.world
The thesis, written as nine short chapters for the BIL conference. Ecology, models, civic life. Simulation as argument. Everything since has been an attempt to build what this essay describes.
@layout: site @image: /anselm/talks/iiw-2026/img/hike.webp @media: /anselm/talks/iiw-2026/img/orbital.webp @section: 2022 @sub: 2022
orbital.foundation
First pass on a digital twin engine. A late-binding event bus, declarative manifests, an entity-component pattern, worked out in Rust first and then JavaScript. Funded by a $320k Futurewei grant.
@layout: site @image: /anselm/talks/iiw-2026/img/volcanoes.webp @media: /anselm/talks/iiw-2026/img/lita.webp @section: A real working hotel @sub: 2025
LITA
A digital twin of a hotel that actually operates, built on Orbital. Guests, staff, robots and lifts. Fire an event (an arrival, a towel, a spill) and watch the building respond.
@layout: site @image: /anselm/talks/iiw-2026/img/reef.webp @media: /anselm/talks/iiw-2026/img/divesight-phones.webp @section: In the App Store @sub: 2025 to 2026 @credit: Tubbataha, Sulu Sea. Photograph by Andrea Marshall, marinemegafauna.org, with permission.
divesight.ai
A digital twin for divers. Marine life and visibility forecasts for 28,000 dive sites, so you know where to dive today. I dive a lot. This one is personal.
@layout: site @image: /anselm/talks/iiw-2026/img/iceberg.webp @media: /anselm/talks/iiw-2026/img/hormuz.webp @section: A playable model @sub: 2026
hormuz.exe.xyz
Close the strait and follow the consequences. Oil finds other ways out; fertiliser, gas and helium do not, and the damage lands in next year's harvests. Every number links to a ledger, and you can change what happens next.
Not entirely happy with this one yet. I think it could communicate better. Work in progress.
@layout: site @image: /anselm/talks/iiw-2026/img/lakevalley.webp @media: /anselm/talks/iiw-2026/img/colorado.webp @section: A playable model @sub: 2026
colorado.exe.xyz
Who drinks the river? The Colorado from source to a delta that used to end in green lagoons. A century of observed flow, Mead and Powell today, and the dashed line of where the curve goes next. You can ask the river.
Also an exploration of computational journalism, or as Bret Victor puts it, explorable explanations. Work in progress.
@layout: site @image: /anselm/talks/iiw-2026/img/crags.webp @media: /anselm/talks/iiw-2026/img/futuresdesk.webp @section: Computational journalism @sub: 2026
futuresdesk.ai
The Future Times. All the news that's fit to model. A newsroom where every story sits on top of a runnable model, so a reader can argue with the numbers instead of the headline. Hormuz and the Colorado are the first two stories.
Same family as the last two. The problem is not only how to write simulations but how to communicate them in a way people care about. Work in progress.
@layout: site @image: /anselm/talks/iiw-2026/img/waterfall.webp @media: /anselm/talks/iiw-2026/img/opencubed.webp @section: A public standard @sub: 2026
opencubed.org
Open³. A working-draft certification for whether a project, product or organization is open on three separate axes: source, data and participation. Twelve criteria, a fixed auditable rubric, every claim backed by public evidence. Self-assessed, evidenced in public, open to challenge.
The mark says only what it says. Version 0.1.
@layout: site @image: /anselm/talks/iiw-2026/img/fitz1.webp @media: /anselm/talks/iiw-2026/img/wiki.webp @section: The map of the space @sub: 2025
digitaltwins.wiki
An atlas of computational representations of the world. Users, use cases, engines, platforms, data, law. An attempt to see the whole field before building more of it.
This one is going well, and there are a few contributors.
@layout: site @image: /anselm/talks/iiw-2026/img/torre.webp @media: /anselm/talks/iiw-2026/img/ontology.webp @section: Now @sub: 2026
ontology.monster
A machine reads the news, works out which real things the articles are about, sorts those things into kinds, and shows how the kinds share parts. Then it puts them on a map, over time. A week of the world, as the objects in it.
@layout: site @image: /anselm/talks/iiw-2026/img/puyehue.webp @media: /anselm/talks/iiw-2026/img/esim.webp @section: Now @sub: 2026
esimv1.exe.xyz
The fresh cut. Terrain, buildings and the fire's arrival in three dimensions, here the 2025 Palisades fire fitted against the record. Modules declare inputs and outputs, ledgers must balance, all data public.
I tend to be tools agnostic, but this simulator is going well. Happy with the initial results, and I may push it further.
@image: /anselm/talks/iiw-2026/img/fitzroy.webp @sub: 2027 and on
Next
- Community. I want to run an event like this one, or build a real-world app for the identity community as a worked example.
- Identity. The critical substrate for mass-market consumer apps that help us navigate what is coming. Everything above sits on it.
I want to help people, and I would like collaborators.
anselm@gmail.com · email, the real one · 415-519-3297 · phone, also real · anselm.substack.com · where I am writing now · anselm.medium.com · where I used to write · x.com/anselm · I retweet a lot of geo stuff · linkedin.com/in/anselm · I repost a ton of simulation and modeling work