# .plan-26-23: Earth Embeddings, Emails Everywhere, and ERRNOOOs

*2026-06-07 — note*


Got some hacking done this week after the [media whirlwind](https://anil.recoil.org/notes/hedgehog-tessera-week), mostly to get [TESSERA](https://anil.recoil.org/projects/tessera) 1.1 out of the door and a stabilisation push on [GeoTessera 0.9](https://github.com/ucam-eo/geotessera).
A quick map of the weekly is [TESSERA v1.1 progress](#tessera-v11-progress) (including [GeoTessera 0.9 onto S3+Zarr](#geotessera-09)); [io-uring and OCaml](#io-uring-and-ocaml); distributed systems covering [self-hosted email](#on-distributed-matters), [Bluesky/standardsite embeds](#bluesky-and-standardsite), [Brent Yorgey on AI and individual choice](#on-ai-and-individual-choice), and [two new Internet-ecology preprints](#a-wild-internet-ecology-enters-the-fray); the kick-off of [Evidence TAP](#evidence-tap-kicks-off); coverage of our [REDD+ work in New Scientist and the WSJ](#carbon-credit-assessments-in-the-new-scientist-and-wsj); book club on [Stone Book Quartet](#the-stone-book-quartet); and [fun links](#fun-links).

## TESSERA v1.1 progress

Hacking continues to get the next iterations of the model and access
libraries out as progress marches on in project [TESSERA](https://anil.recoil.org/projects/tessera)! Some highlights of coverage this week were:

- The [TESSERA v1.0 paper](https://anil.recoil.org/papers/2025-tessera) is at [CVPR 2026](https://anil.recoil.org/notes/cvpr-tessera-esa) in Denver, with [Jovana Knezevic](https://github.com/mahuna13) out there spreading the embeddings gospel\!
- We were on the front page of the [European Space Agency](https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Tessera_AI_model_offers_accessible_way_to_view_Earth), which was incredibly cool. I grabbed a screenshot of ESA's own HQ from the [TZE Explorer](https://tze.geotessera.org) which is the hero image on that piece.
- [phys.org](https://phys.org/news/2026-06-tessera-ai-accessible-view-earth.html) and [Orbital Today](https://orbitaltoday.com/2026/06/05/meet-tessera-the-ai-turning-satellite-images-into-earths-fingerprints/) picked it up too.

### GeoTessera 0.9

I spent a chunk of the week on getting [GeoTessera 0.9 \#278](https://github.com/ucam-eo/geotessera/pull/278) stable. This shifts our embedding downloads over to `s3://tessera-embeddings/` on AWS (moving everything on the download path to be S3-specialised), and adds support for the new [TESSERA v1.1 model](https://huggingface.co/geotessera/TESSERA-V-1.1) variant alongside v1.0. The major improvements in the model are better year-on-year temporal embeddings, and higher quality on tiles with fewer observational passes. 1.1 is looking like a pretty awesome iteration on the base model so far.

Thanks to [Mark Elvers](https://www.tunbury.org/) and [Aneesh Naik](https://aneeshnaik.github.io/) for [help debugging](https://aneeshnaik.github.io/blogposts/20260605_weeknotes_2026_23.html) the performance issues that came out of the shift to Zarr. The issues arise from a lack of concurrency in retrieving HTTP chunks ([zarr-python\#3004](https://github.com/zarr-developers/zarr-python/pull/3004)) and an accidentally [uncompressed coordinate array](https://github.com/ucam-eo/geotessera/pull/281).

I'm a little surprised that AWS is so much slower in HTTP latency than our University server, but I think we don't have edge caching enabled (CloudFront). I'm hoping to push out a release in the next few days so that we can debug performance against that rather than blocking all the users trying to get their [mittens on 1.1 embeddings](https://github.com/ucam-eo/geotessera/issues?q=is%3Aissue%20label%3Aembedding-request).

I did also deploy a quick [v1.1 update to TZE](https://tze.geotessera.org/?store=v1.1) so we can browse the embeddings that we do have, and it results in some very pretty visuals.

<a href="https://tze.geotessera.org/?store=v1.1"> <figure class="image-center"><img src="/images/tze-explorer-v1.1-ss-1.webp" alt="" title="" loading="lazy" srcset="/images/tze-explorer-v1.1-ss-1.768.webp 768w, /images/tze-explorer-v1.1-ss-1.640.webp 640w, /images/tze-explorer-v1.1-ss-1.480.webp 480w, /images/tze-explorer-v1.1-ss-1.320.webp 320w, /images/tze-explorer-v1.1-ss-1.2560.webp 2560w, /images/tze-explorer-v1.1-ss-1.1920.webp 1920w, /images/tze-explorer-v1.1-ss-1.1600.webp 1600w, /images/tze-explorer-v1.1-ss-1.1440.webp 1440w, /images/tze-explorer-v1.1-ss-1.1280.webp 1280w, /images/tze-explorer-v1.1-ss-1.1024.webp 1024w"><figcaption></figcaption></figure> </a>

### io-uring and OCaml

Since [Thomas Leonard](https://github.com/https://roscidus.com) is back and [hacking on Eio](https://notes.roscidus.com/2026/06/01/) I've been doing a refresh of the [ocaml-uring](https://github.com/ocaml-multicore/ocaml-uring) libraries. I need to do a lot of Zarr data copying in the coming weeks and the obvious way to do this is via extremely zero-copy OxCaml code. There's nothing that makes the systems hacker in me happier than having an excuse to write some high performance OCaml code\!

Aside from fixing a bunch of bugs, I've added [shutdown, socket, renameat and symlinkat](https://github.com/ocaml-multicore/ocaml-uring/pull/147), and [fallocate, fsync and ftruncate](https://github.com/ocaml-multicore/ocaml-uring/pull/149), and exposed [Linux-specific errnos](https://github.com/ocaml-multicore/ocaml-uring/pull/152) with hilarious names like [EOWNERDEAD, ENOTRECOVERABLE, ERFKILL, EHWPOISON](https://amok.recoil.org/@avsm/116698423254334531).

I've got another separate tree with OxCaml specific bindings (mostly using
`caml_alloc_local` to go full stack alloc) for my [zero-alloc httpz server](https://anil.recoil.org/notes/oxcaml-httpz) to use, but more on that after I make the server sweat a
little more. [Mark Elvers](https://www.tunbury.org/) has also been [upgrading our CI system](https://www.tunbury.org/2026/06/08/ubuntu-26-04/) to Linux 7.0 so we can test these features more easily, as well as valiantly [trying to keep RISC-V alive](https://www.tunbury.org/2026/06/03/emulated-riscv-workers/) on our nodes despite that architecture's best attempts to become irrelevant through hardware scarcity.

## On distributed matters

I put up a [detailed post on self-hosting email](https://anil.recoil.org/notes/recoil-self-hosting-2026) that's attracting [lively discussion](https://lobste.rs/s/cw7vxa/self_hosting_email_hard_way_from_your_own) on the Interwebs. This came out of the [Rewilding the Web](https://anil.recoil.org/notes/rewilding-the-web-report) workshop I went to last week. I'm going to do a series of blogs on various self-hosting matters over the coming months, as there's clearly appetite for people who want to get control of their own data again.

### Bluesky and standardsite

Aside from email, another important distributed system I use is the 'social database' that underpins Bluesky. One of the promises of this underlying database is that multiple services built over it can interoperate. [Tangled Git hosting](https://anil.recoil.org/notes/disentangling-git-with-bluesky) is one such service, but what else is there?

<figure class="image-right-float"><img src="/images/bsky-standardsite-1.webp" alt="Notice the 'View publication' button below the post that represents the embed" title="Notice the 'View publication' button below the post that represents the embed" loading="lazy" srcset="/images/bsky-standardsite-1.768.webp 768w, /images/bsky-standardsite-1.640.webp 640w, /images/bsky-standardsite-1.480.webp 480w, /images/bsky-standardsite-1.320.webp 320w, /images/bsky-standardsite-1.1024.webp 1024w"><figcaption>Notice the 'View publication' button below the post that represents the embed</figcaption></figure>
Well, this week we saw Bluesky [ship a new feature](https://atproto.com/blog/standard-site-bluesky-timeline) that makes publishing to [standardsite](https://standard.site) records much more worthwhile: the Bluesky client now renders long-form Standard.site document embeds nicely inline. Seeing one of my own `standardsite` documents show up as a proper rich embed in someone else's feed is exactly the kind of small-pieces-loosely-joined interop I keep hoping the network will grow into.

> Standard.site defines a few Lexicons for publishing websites—such as a
> publication (like a website or blog), a document (like an article or post),
> and one for subscriptions (for tracking which publications to follow). Taken
> together, these describe longform writing in an atmospheric way, similar to
> how the Bluesky lexicons describe a social network.
> <cite>\-- [What the Standard.site lexicons do](https://atproto.com/blog/standard-site-bluesky-timeline), June 2026</cite>

Using my [OCaml ATproto](https://tangled.org/anil.recoil.org/ocaml-atp) library, I can automatically add records into the ATProto public relay, and they'll get picked up by Bluesky. For example, this site's record is [browsable here](https://pdsls.dev/at://did:plc:nhyitepp3u4u6fcfboegzcjw/site.standard.publication/3mchoxkwlsx2y):

<a href="https://pdsls.dev/at://did:plc:nhyitepp3u4u6fcfboegzcjw/site.standard.publication/3mchoxkwlsx2y"> <figure class="image-center"><img src="/images/pdsls-dev-1.webp" alt="" title="" loading="lazy" srcset="/images/pdsls-dev-1.768.webp 768w, /images/pdsls-dev-1.640.webp 640w, /images/pdsls-dev-1.480.webp 480w, /images/pdsls-dev-1.320.webp 320w, /images/pdsls-dev-1.1024.webp 1024w"><figcaption></figcaption></figure> </a>

### On AI and individual choice

There's an excellent essay by Brent Yorgey called "[*To my students*](http://ozark.hendrix.edu/~yorgey/forest/00FD/)" on how to carry yourself through a software industry that's being entirely disrupted by AI at the moment. It's written as advice to his compsci students, and it's helped me form up my thoughts around [AI disclosure proposal](https://anil.recoil.org/notes/opam-ai-disclosure):

> Don't believe self-serving lies about technologies being "inevitable" or
> "here to stay". You don't have to just go along with the dominant narrative.
> You can make deliberate choices and help others to do the same.
> <cite>\-- [Brent Yorgey, *"To my students"*](http://ozark.hendrix.edu/~yorgey/forest/00FD/), 2026</cite>


He concludes with "be motivated by love instead of fear", which is a bloody brilliant way to think about the current choices we all face. My view on self hosting and open source is that I want to preserve the right for future generations to have their own agency about how digital technologies will guide their lives. I'm entirely for the use of AI *if all parties involved are informed and consenting*, which is what motivated my [disclosure proposal](https://anil.recoil.org/notes/opam-ai-disclosure-update) the other week.

It's also great to see another OCaml-powered [Forester](https://www.forester-notes.org/index/index.xml) blog spring up in the wild. Well done on growing a userbase with such excellently thoughtful content, [Jon Sterling](https://jonmsterling.com)\!

### A wild Internet ecology enters the fray

A couple of preprints popped up this week that are relevant to the [internet ecology](https://anil.recoil.org/papers/2025-internet-ecology) work I've been seeding on [antibotty networks](https://anil.recoil.org/notes/ecology-at-aarhus) and self-modifying code:

- [AI Agents Enable Adaptive Computer Worms](https://arxiv.org/abs/2606.03811) from Nicolas Papernot's lab (with Cambridge's own Hanna Foerster among the authors) demonstrates a worm that reasons about each target and decides how to attack it using an open-weight LLM model. Like Morris' original worm, it parasitically runs the LLM on the machines it compromises.  It was covered in the [New York Times](https://www.nytimes.com/2026/06/02/technology/scientists-find-way-to-supercharge-dangerous-computer-worms-with-ai.html) as well as [heise](https://www.heise.de/en/news/IT-researchers-demonstrate-adaptive-AI-worm-11318259.html).
- [ExploitGym](https://arxiv.org/abs/2605.11086) is a benchmark of ~900 real-world vulnerability instances for measuring whether AI agents can turn theoretical vulnerabilities into working exploits. They showed, somewhat worryingly, that the three major frontier models all found *different* sets of exploits in the same code. There's a lot of latent bugginess, which hopefully can be turned into [cryptic resilience](https://doi.org/10.1098/rspb.2025.2377) rather than pure vulnerability\!

## Evidence TAP kicks off

After [playing with LEGO](https://anil.recoil.org/notes/2026w7) earlier in the year, we had the first meeting of our new Evidence TAP project, which is a broadening of [Conservation Evidence](https://anil.recoil.org/projects/ce) into new fields such as education. On the education side, we had [Jenny Gibson](https://www.educ.cam.ac.uk/people/staff/gibson/) and Mélanie Gréaux who has just [returned after her PhD](https://www.languagesciences.cam.ac.uk/news/researcher-profile-melanie-greaux) to come back and work on education evidence synthesis fulltime with us! [Welcome Mel, back to Cambridge\!](https://www.linkedin.com/feed/update/urn:li:activity:7465687486401265664/):

> In Cambridge, I am joining a fantastic multidisciplinary team working on an
> ambitious project to build an AI-facilitated evidence platform helping
> policymakers and practitioners to find evidence-based answers to their
> questions: what works, for whom, and in which contexts? In the age of AI, we
> are presented with an opportunity to re-think our approach and access to
> evidence systems. If done well, it can accelerate the democratisation of
> knowledge and strengthen our ethical engagement with data. I’ll be leading
> work on evidence systems for early childhood education.
> <cite>\-- [Mélanie Gréaux](https://www.linkedin.com/feed/update/urn:li:activity:7465687486401265664/), LinkedIn, 2026</cite>

It was very cool to traipse over to the [Donald Mcintyre Building](https://commons.wikimedia.org/wiki/File:Donald_Mcintyre_Building,_Faculty_of_Education,_Cambridge.jpg) beside Homerton College with [Eleanor Toye Scott](https://www.cst.cam.ac.uk/people/eft20), [Sadiq Jaffer](https://toao.com), [Lynn Dicks](https://www.zoo.cam.ac.uk/directory/prof-lynn-dicks), Rob Doubleday and [Bill Sutherland](https://www.zoo.cam.ac.uk/directory/bill-sutherland) and see the inside of a building I hadn't seen before after many years in Cambridge\!

More on the Evidence TAP in the coming weeks; I need to knock up a website for the project next week to give us a blogging area!  Having three departments involved means that we need more coordination than usual, but our [Zulip](https://eeg.zulipchat.com) is serving well so far.

<figure class="image-center"><img src="/images/evitap-education-1.webp" alt="In the Donald Mcintyre building for the first time (for me)!" title="In the Donald Mcintyre building for the first time (for me)!" loading="lazy" srcset="/images/evitap-education-1.768.webp 768w, /images/evitap-education-1.640.webp 640w, /images/evitap-education-1.480.webp 480w, /images/evitap-education-1.3840.webp 3840w, /images/evitap-education-1.320.webp 320w, /images/evitap-education-1.2560.webp 2560w, /images/evitap-education-1.1920.webp 1920w, /images/evitap-education-1.1600.webp 1600w, /images/evitap-education-1.1440.webp 1440w, /images/evitap-education-1.1280.webp 1280w, /images/evitap-education-1.1024.webp 1024w"><figcaption>In the Donald Mcintyre building for the first time (for me)!</figcaption></figure>

## Carbon credit assessments in the New Scientist and WSJ

<a href="https://www.newscientist.com/article/2525921-carbon-credits-are-flawed-but-they-can-still-help-save-forests/"> <figure class="image-right-float"><img src="/images/new-scientist-cc-ss-1.webp" alt="" title="" loading="lazy" srcset="/images/new-scientist-cc-ss-1.768.webp 768w, /images/new-scientist-cc-ss-1.640.webp 640w, /images/new-scientist-cc-ss-1.480.webp 480w, /images/new-scientist-cc-ss-1.320.webp 320w, /images/new-scientist-cc-ss-1.1600.webp 1600w, /images/new-scientist-cc-ss-1.1440.webp 1440w, /images/new-scientist-cc-ss-1.1280.webp 1280w, /images/new-scientist-cc-ss-1.1024.webp 1024w"><figcaption></figcaption></figure> </a>
I missed this a few weeks ago, but it turns out we had a couple of articles covering our research [on REDD+ overcrediting](https://anil.recoil.org/notes/redd-overcrediting) in both New Scientist and the Wall Street Journal! The paper involved was "[Learning lessons from over-crediting to ensure additionality in forest carbon credits](https://anil.recoil.org/papers/2025-redd-evals)".

First, New Scientist had a nice piece on "[Carbon credits are flawed, but they can still help save forests](https://www.newscientist.com/article/2525921-carbon-credits-are-flawed-but-they-can-still-help-save-forests/)" ([archive.is mirror](https://archive.is/TfRRu)):

> Carbon credits bought by companies to offset their emissions really have
> reduced deforestation, but not by as much as credit developers claim, according
> to a rigorous analysis.
> 
> \[...\] So who's right? Both, according to a growing body of research. Last
> month, one of the most rigorous studies yet found that most early projects
> did successfully reduce deforestation. But they sold credits for almost 11
> times more forest on average than they actually saved.
> <cite>\-- [New Scientist](https://www.newscientist.com/article/2525921-carbon-credits-are-flawed-but-they-can-still-help-save-forests/)</cite>

Then the Wall Street Journal also had a piece on "[The Way Companies Aim for Net Zero Is Flawed. It's Also Working](https://www.wsj.com/pro/sustainable-business/the-way-companies-aim-for-net-zero-is-flawed-its-also-working-07099d0e)" ([archive.is mirror](https://archive.is/stF0V)):

> Two academic papers published last month show corporate climate efforts
> having a positive impact on reducing deforestation and cutting emissions.
> 
> \[...\] "What is really positive about carbon credits is that despite being a
> complicated economic instrument for achieving these outcomes, they do at
> least make it possible for us to directly fund efforts that reduce
> deforestation on the ground," said Tom Swinfield, one of the report's
> authors. "They allow you to circumvent all the complex politics and get to
> the heart of the problem."
> <cite>[The Way Companies Aim for Net Zero Is Flawed. It’s Also Working](https://www.wsj.com/pro/sustainable-business/the-way-companies-aim-for-net-zero-is-flawed-its-also-working-07099d0e), WSJ, May 2026</cite>

 
## The Stone Book Quartet

In our Christ's book club organised by [Jenny Gibson](https://www.educ.cam.ac.uk/people/staff/gibson/), we had [Richard Mortier](https://github.com/mor1) recommend Alan Garner's [Stone Book Quartet](https://en.wikipedia.org/wiki/The_Stone_Book_Quartet). I did struggle with the Cheshire dialect in his prose, but absolutely loved the opening of the first "The Stone Book", where a child climbs to the top of a steeple and is then taken deep underground to be shown a cave painting in an opening so narrow only a child can crawl through.

I always get the bug to go and visit the places in the books I've read, like trekking out to Dartmouth after finishing Julian May's [Galactic Milieu saga](https://en.wikipedia.org/wiki/Galactic_Milieu), and after this one I very much want to go and walk [Alderley Edge](https://en.wikipedia.org/wiki/Alderley_Edge), the bit of Cheshire that runs all the way through Garner's writing and that has had human settlements for a very loong time.

## Fun Links

- [Justin Cormack](https://github.com/justincormack) recommends [*The Irrational Decision: How We Gave Computers the Power to Choose for Us*](https://press.princeton.edu/books/hardcover/9780691272443/the-irrational-decision) by Ben Recht, on how a 1940s statistical definition of rationality became the dominant framework for decision-making everywhere.
- Lovely little library from Aneesh Naik that I want to use called [datahues](https://github.com/aneeshnaik/datahues) which generates perceptually-uniform colour ramps by interpolating in [Oklab](https://bottosson.github.io/posts/oklab/) space, so you don't see banding patterns in your data that aren't really there. Seems perfect for
Synopsis: TESSERA on the ESA homepage and at CVPR, GeoTessera 0.9 stabilising onto S3/Zarr, io-uring in OCaml, carbon credits in New Scientist and WSJ, and musings on internet malware again.
Words: 2010
DOI: 10.59350/972b1-n4440

## Related

- [Tessera at CVPR 2026, and the front page of the European Space Agency!](https://anil.recoil.org/notes/cvpr-tessera-esa) (note, 2026-06-06)
- [Self-hosting email the hard way from your own routable IPv4 block up](https://anil.recoil.org/notes/recoil-self-hosting-2026) (note, 2026-06-06)
- [Rewilding the Web: my workshop report from Edinburgh](https://anil.recoil.org/notes/rewilding-the-web-report) (note, 2026-05-30)
- [What happens when a hedgehog story prickles its way into the BBC](https://anil.recoil.org/notes/hedgehog-tessera-week) (note, 2026-05-22)
- [Voluntary AI disclosure proposal for OCaml: update 1](https://anil.recoil.org/notes/opam-ai-disclosure-update) (note, 2026-05-14)
- [Helping tropical forest protection keep up with a fast-changing world](https://anil.recoil.org/notes/redd-overcrediting) (note, 2026-04-30)
- [A Proposal for Voluntary AI Disclosure in OCaml Code](https://anil.recoil.org/notes/opam-ai-disclosure) (note, 2026-04-03)
- [Learning lessons from over-crediting to ensure additionality in forest carbon credits](https://anil.recoil.org/papers/2025-redd-evals) (paper, 2026-04-01)
- [.plan-26-07: Storage, Lego, Echo, and the IUCN](https://anil.recoil.org/notes/2026w7) (note, 2026-02-15)
- [My (very) fast zero-allocation webserver using OxCaml](https://anil.recoil.org/notes/oxcaml-httpz) (note, 2026-02-01)
- [TESSERA: Temporal Embeddings of Surface Spectra for Earth Representation and Analysis](https://anil.recoil.org/papers/2025-tessera) (paper, 2025-11-01)
- [Presenting our Ecology of the Internet ideas at Aarhus 2025](https://anil.recoil.org/notes/ecology-at-aarhus) (note, 2025-08-22)
- [Steps towards an Ecology for the Internet](https://anil.recoil.org/papers/2025-internet-ecology) (paper, 2025-08-01)
- [Socially self-hosting source code with Tangled on Bluesky](https://anil.recoil.org/notes/disentangling-git-with-bluesky) (note, 2025-03-08)
- [TESSERA, a pixelwise geospatial foundation model](https://anil.recoil.org/projects/tessera) (project, 2025-01-01)
- [Conservation Evidence Copilots](https://anil.recoil.org/projects/ce) (project, 2024-01-01)

---
Canonical: https://anil.recoil.org/notes/2026w23
Type: note
Tags: reading, tessera, networking, zarr, ai, standards, philosophy