Darren Wilkinson
banner
darrenjw.bsky.social
Darren Wilkinson
@darrenjw.bsky.social
Professor of Statistics, Bayesian, computational systems biologist and functional programmer, https://darrenjw.github.io/
Reposted by Darren Wilkinson
Hot off the press - a letter from the Exec Chair of @ukri.org explaining his plans and in particular the financial position of STFC. It notes curiosity driven research is protected and will make up around half of UKRI spend over the coming period
www.ukri.org/news/open-le...
Open letter from Ian Chapman to research and innovation community
UKRI Chief Executive outlines changes to UKRI investment approach, addressing concerns about research funding and the financial position of STFC.
www.ukri.org
February 1, 2026 at 5:02 PM
Reposted by Darren Wilkinson
New open source: cuthbert πŸ›

State space models with all the hotness: (temporally) parallelisable, JAX, Kalman, SMC
January 30, 2026 at 4:26 PM
Reposted by Darren Wilkinson
The first data science book that has a chapter on monads reproducible-data-science.dev

Learn how to build robust #DataScience pipelines with #RStats, #Python , #Julia and #Nix !
February 1, 2026 at 11:47 AM
Reposted by Darren Wilkinson
BBSRC and EPSRC join MRC "pausing" funding grant apps www.timeshighereducation.com/news/three-u...
Three UK research councils suspend funding opportunities
Temporary blockΒ on grant applications by MRC, BBSRC and EPSRC heightens concerns over funding cuts
www.timeshighereducation.com
January 31, 2026 at 9:38 AM
Reposted by Darren Wilkinson
Hivemind question: Palmer Penguins is doing great work avoiding overuse of Fisher's iris data. What other "dataset substitutions" do you like? Or just generally other cool, real, open datasets that are good for teaching because established methods work pretty well on them?
January 4, 2026 at 3:43 PM
It's not something I've looked into deeply, but I do look at the close relationship between discrete and continuous (space and) time in these notes, so you may find them relevant.
January 1, 2026 at 11:05 PM
Reposted by Darren Wilkinson
I made a (sort of) "Tinder for Git" site to help developers find 'good first issues' in open-source projects. It's early days, but it could be helpful.

gitafix.com/
January 1, 2026 at 10:39 PM
I used to teach a course quite close to this. Notes here: darrenjw.github.io/work/teachin...
darrenjw.github.io
January 1, 2026 at 10:33 PM
LOL! Imperative programmers are just determined to burn the whole world down! 😜
January 1, 2026 at 9:46 PM
Some lecture notes on spatial and temporal modelling, illustrated using R - darrenjw.github.io/spatio-tempo... - #rstats #rspatial #quarto
Spatial and temporal statistics
darrenjw.github.io
January 1, 2026 at 12:31 PM
The "netcontrol" #rstats package was recently removed from #cran - I have some instructions for installing the archived version: darrenjw.github.io/work/docs/ne...
netcontrol for R
darrenjw.github.io
December 5, 2025 at 1:51 PM
Stochastic Gray Scott diffusion via the spatial chemical Langevin equation.
November 27, 2025 at 9:20 PM
The Pennine Way is quite beginner-friendly, and intersects Hadrian's wall. en.wikipedia.org/wiki/Pennine...
Pennine Way - Wikipedia
en.wikipedia.org
November 7, 2025 at 10:33 PM
Reposted by Darren Wilkinson
I'll admit, I was skeptical when they said Gemini was just like a bunch of PhDs. But I gotta admit they nailed it.
August 17, 2025 at 1:51 PM
Reposted by Darren Wilkinson
Staff fear UK's Turing AI Institute at risk of collapse www.bbc.com/news/article...

BBC News has seen the complaint that the Guardian reported on Sunday. Main additional info seems to be that Helen Margetts and Cosmina Dorobantu left positions at the ATI last month.

#govtech #techpolicy #AIpolicy
August 12, 2025 at 5:35 AM
"ChatGPT will apologize for anything - even advice it definitely didn't give". Is ChatGPT... British?!
"I'm sorry for advising you to accept the trade of your cow for three beans. In hindsight, that deal was not supported by conventional livestock market values, nor was it backed by any credible bean-based economic index."

www.aiweirdness.com/chatgpt-will...
ChatGPT will apologize for anything
ChatGPT will apologize for anything - even advice it definitely didn't give, and stuff it definitely didn't do. It very much regrets its recommendation that we hire a giraffe as CEO.
www.aiweirdness.com
August 8, 2025 at 5:17 PM
Little summer holiday project: learn enough about fluid dynamics to implement a 2d simulation from scratch: www.youtube.com/watch?v=DYos...
2d incompressible fluid flow
YouTube video by Darren Wilkinson
www.youtube.com
August 7, 2025 at 5:44 PM
Probably two different issues at play here. Turing hero- worshiping is probably a bit nationalistic, but I think the call for spending on defense related research is more a reflection of the current geopolitical landscape.
July 21, 2025 at 4:33 PM
Yes. There are a bunch of issues with the Turing, but it's not clear (to me) that this is a better path forward.
July 21, 2025 at 4:20 PM
Reposted by Darren Wilkinson
Going to the hospital because I broke my wrist smashing the endorse button:
www.understandingai.org/p/i-got-fool...
I got fooled by AI-for-science hypeβ€”here's what it taught me
I used AI in my plasma physics research and it didn’t go the way I expected.
www.understandingai.org
May 19, 2025 at 6:04 PM
Yes, Haskell is great, but as well as the critical mass issue that Scala has, it has the additional, more fundamental issue that I don't think lazy languages are ideal for scientific and statistical computing (controversial, I know).
July 5, 2025 at 8:49 AM
Reposted by Darren Wilkinson
Reflections on Haskell and Rust. ~ Sibi Prabakaran. academy.fpblock.com/blog/rust-ha... #Haskell #FunctionalProgramming #Rust
Reflections on Haskell and Rust
Haskell vs. Rust: A production-level comparison of two languages.
academy.fpblock.com
July 4, 2025 at 2:28 PM
Actually, I think I misinterpreted your question. I much prefer strongly typed functional languages (especially Scala 3), but although Scala is popular for big data analytics, there isn't really a critical mass of people using it for serious statistical computing and machine learning.
July 4, 2025 at 2:15 PM
Let's just say that they both have their areas of strength and weakness. But the point I make in the podcast is that they are actually both very similar when considered in the space of all programming languages.
July 4, 2025 at 1:10 PM