With the vast amount of data we have these days for our planetary computing processing, it's naturally tempting to use more hardware offload. The obvious choice, GPGPUs, are not a great fit for the problem due to the difficulty of unlocking high data parallelism for geospatial data. So it's back to an old technology I worked on twelve years ago in the form of FPGAs!

We've been happy users of Matrix for our group communications in the EEG. Today we've been bringing in more members of the wider group to using it instead of Slack. As part of that, I've set up a cool bot called Hookshot which allows Matrix to be connected to external services such as GitHub and Atom/RSS feeds. This is a test post to demonstrate to the members of the EEG how Matrix and Atom work!

Sadiq Jaffer sent along this piece in The Conversation last week about the remarkable number of academic papers that are now AI generated. The numbers of these papers are probably underestimated: What caught my eye in this article is their development of the Problematic Paper Screener, which the good folks at Retraction Watch developed. It works with high precision to detect papers issued by grammar-based generators.

There's a letter in Science today from a bunch of well known remote sensing researchers that make the unusual point that modern satellite resolution is getting too good to be accurate for forest carbon estimation. The lead author Laura Duncanson is a remote sensing scientist at Maryland who works on the incredible GEDI instrument on the International Space Station.

The terms carbon credits and carbon offsets are often used interchangeably, but are in fact two distinct concepts. I've spent a nice Sunday morning reading up on some recent articles that Bhaskar Vira sent me which introduce a third term, known as "carbon contributions" . Rather than this adding confusion, I found it helped me clarify my own thoughts on the matter, which I note down here in draft form.

Affordable digitisation of insect collections using photogrammetry This is an idea proposed in 2025 as a Cambridge Computer Science Part III or MPhil project, and is currently being worked on by Beatrice Spence, Arissa-Elena Rotunjanu and Anna Yiu. It is co-supervised with Tiffany Ki and Edgar Turner.

Srinivasan Keshav posted a link to this excellent deepdive by Prasad Raje of Udemy into the advances that DeepSeek R1 has made from a perspective of the core technology.

Using computational SSDs for vector databases This is an idea proposed in 2025 as a Cambridge Computer Science Part III or MPhil project, and is available for being worked on. It may be co-supervised with Sadiq Jaffer. Large pre-trained models can be used to embed media/documents into concise vector representations with the property that vectors that are "close" to each other are semantically related.

Now that I've switched to a new website, I'm working on open-sourcing its components. I've got a lot of small OCaml scripts that are all work-in-progress, and so not quite suitable to be published to the central opam-repository but I still need be able to run them conveniently on my own self-hosted infrastructure.

My colleagues Thomas Swinfield and Eleanor Toye Scott lead the publication of a comprehensive report of the steps the voluntary carbon market needs to take to restore its scientific credibility, with input from many of us in 4C and beyond. This paper represents three years of hard work from the team on trying to blend remote sensing with carbon quantification.

This website has been through quite a few iterations over the years. The first version in 1998 was written in Perl and hosted on OpenBSD; the second was rewritten in 2000 when I got commit access to PHP; the third rewrite became a hybrid OCaml/PHP/Perl special in 2004 in Blosxom; then the forth rewrite around 2013 got turned into a unikernel in MirageOS;