Aron ⇋ e⁻

A wild digital ride from playing Tetris on a Gameboy as a child to modifying a SOUL.md file that shapes the identity of an autonomous agent

This robot must have some stories to tell #Londonmaxxing

A timely and eloquent essay, well worth reading by anyone in science.

If you've got something really big - 100,000 atoms, a million atoms? I'd really like to try it out.

Come work with me on generative models for materials at CuspAI! We have: - a fantastic team lead by @wellingmax.bsky.social and @aronwalsh.github.io - important problems and fun methods - offices in beautiful Amsterdam, Cambridge, London, Berlin jobs.ashbyhq.com/cuspai/b8108...

Finally made it to #CCSC2026! Amazing talks this morning (day 3) from @vlderinger.bsky.social @mahrossi.bsky.social @molecularxtal.bsky.social 🤩 great to be here

#CompChem for life

The Next Generation of AI Chemists: a programme for penultimate-year UK-based undergrads from underrepresented backgrounds (women, gender minorities, LGBTQ+, ethnic minority, disability, carers, low socioeconomic background, first in family to uni). #chemsky aichemy.ac.uk/next-generat...

Day 🔟 of the NRMP 5th Anniversary Highlights 🧁 Last, but not least, we chatted with @aronwalsh.github.io to learn more about the trends, opportunities, and challenges in using #AI for material design. 🌟 www.nature.com/articles/s43...

Defect simulations are notoriously sensitive to the many choices required 👨‍💻📊 In this Perspective, we highlight best practices in calculating 𝘢𝘯𝘥 𝘳𝘦𝘱𝘰𝘳𝘵𝘪𝘯𝘨 defect properties, hoping to establish guidelines for robust, transparent and reproducible defect simulations 🌟 www.nature.com/articles/s41...

So happy, new chalk drawing in my lecture room this morning! 🤩 Thanks www.instagram.com/lestableauxd...

"Now What?" by John J Hopfield is a beautiful read for researchers at all career stages, and especially so for those who wish not to simply solve problems, but rather identify and solve a PROBLEM -- a kind of the deepest question. pni.princeton.edu/sites/g/file...

A winter refresh in Porto #BraggYourPattern

Come work with us at @cuspai.bsky.social in the generative model team! Excited about flow / diffusion models and chemistry? Looking for impact? jobs.ashbyhq.com/cuspai/b8108... Join a great team lead by @wellingmax.bsky.social and Aron Walsh, work in Amsterdam / Cambridge / London / Berlin. 1/2

Only four decades late to the party, but The Sugarcubes are great. Soundtrack for an afternoon of writing, while watching the rain pour outside www.youtube.com/watch?v=vz9W...

How could you not love #crystallography? An approximation of the incommensurate structure of Y⅓TaO₃ #ChemSky arxiv.org/pdf/2511.01241

Solar capacity is expected to triple based on new projects in the pipeline. Total global solar capacity, including all projects in construction and planned, is expected to reach almost 3 terawatts, with 80% concentrated in just 15 countries. Data @globalenergymon.bsky.social

"Your invitation to review for..." x n, where n is a number that's larger than you expect #ChemSky

"We have successfully received your letter of recommendation" x n, where n is a number that's larger than you expect #ChemSky

Computer scientist Yoshua Bengio has become the first person to have their work cited more than one million times on the search engine Google Scholar go.nature.com/4oywvUV

I'm hiring postdocs @berkeleylab.lbl.gov to drive cutting-edge research involving MLIPs, high-throughput workflows, chemical reaction networks, generative models, and open-source software dev. Full position description + application here: forms.gle/zePBZDmciXez... #Chempostdoc #AI4Science

👋New preprint: "Interplay of Redox Non-Innocence and Symmetry Breaking in a 4d Coordination Framework" 🧪We’ve made Mo(pyz)₂I₂, the 4d cousin of the Cr(pyz)₂X₂ family — with redox-active ligands, symmetry breaking, semiconducting transport, and very strong magnetic interactions. 🔗👇

arXiv will no longer accept review articles and position papers unless they have been accepted at a journal or a conference and complete successful peer review. This is due to being overwhelmed by a hundreds of AI generated papers a month. Yet another open submission process killed by LLMs.

I am not sure how many solid-state chemists are lurking here, but I bet 10p that CaSnO3 is not magnetic (with or without CO2) pubs.rsc.org/en/Content/A...

🤨No & No again! 👉 "Stale Citations" are an integral part of science: to give credit to crucial works of the past, avoiding 'reinventions', being a plague in research, e.g. in #chemsky #pisky Your biased AI Tool to keep "research fresh" is a tombstone of science. Spooky indeed 😱 #researchintegrity

So much software reuse is untracked as people cite the software in the SI or because there are upper limits to citations. 🙃 I have never understood this.

Imagine if Alan Turing had access to Claude Code #CompChem

Axel Becke, one of the big names in DFT development over the past 40 years, has passed away #CompChem #ChemSky (source Russell Boyd)

Next Thursday afternoon, 30/10, the @thomasyoungcentre.bsky.social is organising a symposium in London on modelling battery materials with as speakers Marie-Liesse Doublet (Montpellier), Alex Squires (Birmingham) and Benjamin Morgan (Bath).

The number of emails from potential PhDs & postdocs has skyrocketed this year. I can see many factors (from geopoliticial to scientific), but it's genuinely tough to deal with. I also remember the feeling of getting no response... #ChemSky

Lots of great chemsky content in @science.org this week, including a remarkable porphyrin nanobelt from the Saywell and Anderson groups 🧪 www.science.org/doi/10.1126/...

Most journals now accept format-free submission (and the others should). Another example of “using AI” to solve a problem that could easily be solved another way?

🔥We're excited to announce a major milestone for the machine-learned interatomic potential (MLIP) ecosystem: TorchSim is moving to community ownership and governance through a partnership with Radical AI and the open-source community! TorchSim is an atomistic simulation engine built for the AI era.

In case you haven't seen it yet, the 2023 film Rye Lane is on the BBC iPlayer and is a total delight ▶️📺 www.bbc.co.uk/iplayer/epis...

Historical evidence of life before matplotlib #CompChem

my PhD student G. Fabusola trained and tested machine learning algorithms to parse the response pattern of a conductive-MOF sensor array from K. Mirica's group! 👃 the electronic nose could detect and differentiate toxic gases and H₂S/SO₂ mixtures at ppm-levels. pubs.acs.org/doi/10.1021/...

You never know what you’ll find while wandering around London #ChemSky

Very strange to see #MOFs on mainstream tv after the #NobelPrize in #chemistry announcement. I guess it’s going to pop up in everyday lives more than usual! Congratulations again to the winners and a special shoutout to Prof. Susumu Kitagawa - a partial alumni gathering photo below #chemsky

Just a little humble brag. Here's me getting my PhD with my supervisors - Bernard Hoskins and newly minted #NobelPrize Laureate Richard Robson. So thrilled for Richard, and thinking also of Bernard tonight (who passed away many years ago but was a crucial collaborator of Richard's). #Chemsky #ozchem

"We propose that a new and potentially extensive class of solid polymeric materials with unprecedented and possibly useful properties..." Quite a strong opening line by Hoskins and Robson pubs.acs.org/doi/abs/10.1...

Not sure this was intended as a materials discovery sign... but it is now #chemsky #compchemsky

Yeah. Sorry. Also: I hope you'll buy my books.

"Review articles are a relic of the past" - overheard in a meeting yesterday. Deep research tools give a focused & referenced report to your specification. @kamat.bsky.social wrote a nice piece (pubs.acs.org/doi/10.1021/...), which emphasises that the key ingredient is saying something new #ChemSky

My tutorial review on single crystal diffraction analysis is now online @crystengcomm.rsc.org! Students, supervisors, and reviewers: Please consider this thought-provoking read to discover many facets of the crystallographic practice that are often overlooked or misunderstood 💎 #crystallography

I've complained a couple times about the lack of physics community here but not done much about it! New commitment: I'll post one interesting arXiv paper a day until things are better. (To the folks who follow me for non-physics science policy, don't worry; plenty more of that reporting to come.)

Defect calculations have many pitfalls and key considerations for achieving good accuracy 🎯 In this perspective, we discuss these issues, how to avoid and how we can make defect simulations more reproducible – particularly important with more ML developments! 📊 chemrxiv.org/engage/chemr...

Celebrated the day 1 of my new hobby with this coolest framework