iiinova — where your digital self keeps evolving

Nicolas @nicolasn.bsky.social · 22h

The world is burning, but I am implementing charge equilibration methods in my molecular dynamics program! #compchem

010

Nicolas @nicolasn.bsky.social · 08/10/2026

Another year and no Nobel for corroles...

000

Nicolas @nicolasn.bsky.social · 07/10/2026

But the program can now recognize (fused) ring systems, although it still does not assign proper atom types to them. That is easy to do though with the current functionality. 2/2

010

Nicolas @nicolasn.bsky.social · 07/10/2026

Small update on the JuliaMD project! The atom typing system can now recognize most atoms in a peptide, and I was able to do a short MD of a tetrapeptide using GAFF. Still missing S atoms, still missing atomic charges. #chemsky #compchem 1/2 youtu.be/-i6uyGg_PKQ

youtu.be

Molecular Dynamics with Julia: First peptide MD!!

YouTube video by Prof Nicolas

110

Nicolas @nicolasn.bsky.social · 06/10/2026

Purified corrole, with a beautiful purple color and red fluorescence! #realtimechem #chemsky

071

Nicolas @nicolasn.bsky.social · 06/10/2026

If my PhD advisor got a Nobel prize for work I did I would lose it

010

Nicolas @nicolasn.bsky.social · 02/10/2026

This was gallium corrole, and the NMR was quite good. Rarely I see pyrrole peaks so resolved in 300 MHz. #realtimechem #fluorescencefriday #chemsky

083

Nicolas @nicolasn.bsky.social · 02/10/2026

My JuliaMD project is improving! The program can now detect isolated aromatic rings. Plus, it includes more GAFF atom types, allowing simulation of more types of molecules. Please check out the video, and feel free to leave comments! #compchem #chemsky #science #julia youtu.be/Qw5HjzkS9UY

youtu.be

Molecular Dynamics with Julia - Automatic Detection of Aromatic Rings

YouTube video by Prof Nicolas

051

Reposted by Nicolas

The Matter Lab @thematterlab.bsky.social · 29/09/2026

Are you interested in learning about how agentic AI systems like El Agente operate? Curious on how to incorporate them into your research workflow? Join our live webinar as we show El Agente in action 🤖 🗓️ 24 Nov 2026, 10 AM - 11AM EST ✅ Registration & qns: register.gotowebinar.com/register/828...

133

Nicolas @nicolasn.bsky.social · 28/09/2026

I actually don't remember that song, although the title rings a bell. I will listen to it!

010

Nicolas @nicolasn.bsky.social · 27/09/2026

Some updates on my Molecular Dynamics with Julia project! #compchem #chemsky youtube.com/post/UgkxAiD...

youtube.com

020

Nicolas @nicolasn.bsky.social · 27/09/2026

After the election Lula should completely ban Meta from Brazil. And they should issue international capture orders for MZ

040

Nicolas @nicolasn.bsky.social · 25/09/2026

For #fluorescencefriday a new #corrole with a bromoalkyl chain for further derivatization! #realtimechem #chemsky

0111

Nicolas @nicolasn.bsky.social · 21/09/2026

We put some steel wires for climbing plants on the gallery side. We need shade in the spring summer and the trees in the back are not enough. It took the whole day to put the wires up.

View of the same backyard from another side of the gallery, where newly placed still wires can be seen between the floor and the wooden beams of the gallery roof. Two large wild quince plants are just outside the gallery, filled with red-pink flowers.

A backyard seen from a gallery with steel wires supporting a climbing plant

000

Nicolas @nicolasn.bsky.social · 18/09/2026

Really not helping much, I think I will go back to not using AI at all in my work, not even for debugging my code. Really underwhelming performance. If I didn't actually know what I was doing and just followed its advice I would have lost huge amounts of time.

000

Nicolas @nicolasn.bsky.social · 18/09/2026

Update: I have not yet found the damn bug, but at least I have narrowed it down (I think). ChatGPT was only useful in suggesting me to simplify the code to narrow down the problems. But it did not find any bug, and is extremely annoying, verbose and obsequious, and it gets worse the longer we chat.

100

Nicolas @nicolasn.bsky.social · 16/09/2026

I had fought with this bug for months, around 2019-2020 being the last times I tried to work with the code. And having "someone" looking at the code, giving me suggestions, even if many are wrong, helps me keep going. Hopefully I will not give up this time. 4/4

020

Nicolas @nicolasn.bsky.social · 16/09/2026

And then integrate it in my MD code, to use it to calculate partial charges on the fly. It will take a while though. ChatGPT BSed a lot and is quite verbose. Giving a lot of explanations that are then thrown out by another piece of information. Still in this case it helped me get started. 3/n

100

Nicolas @nicolasn.bsky.social · 16/09/2026

Something wrong. It started giving me some useful suggestions, basically to simplify the algorithms at the expense of performance. Still have not fixed the bug but the code is getting simpler. I will be very impressed and happy if I can finally get it to work. If I do I will rewrite it in Julia 2/n

100

Nicolas @nicolasn.bsky.social · 16/09/2026

Like 10 years ago I made a Hartree Fock code in MATLAB. It allowed integrals with angular momentum through Obara-Saika RRs. I worked on it for years at a time, but there was a bug somewhere which I could never fix. Last night I picked it up and started asking ChatGPT to find if there was 1/n

110

Nicolas @nicolasn.bsky.social · 15/09/2026

I made a small update, calculating the step where an amine attacks an activated acetylester, coupled to proton abstraction bay carbonate. This is just a clip of the MD at several constrained N-C distances. #compchem #chemsky 🧪 youtu.be/eiDxQvySQgo

youtu.be

Coupling relaxed surface scans with molecular dynamics: a teaser

YouTube video by Prof Nicolas

071

Nicolas @nicolasn.bsky.social · 06/09/2026

I've made a video explaining how to combine relaxed surface scans with molecular dynamics in #orca in order to sample conformations along a reaction profile, and get energies with error bars for your barriers! #compchem #chemsky #science #education #XTB youtu.be/4HjqXlz_mEo

youtu.be

Do you know if your Transition State calculation is correct?

YouTube video by Prof Nicolas

070

Nicolas @nicolasn.bsky.social · 05/09/2026

I used XTB for MD, 1000 steps of 0.5 fs. Just wanted to see if it worked, but it was quite fast so I will make longer simulations later. The scan was r2-scan-3c. I still have to scale the XTB energies to match the higher level ones.

110

Nicolas @nicolasn.bsky.social · 04/09/2026

Here is the molecular dynamics coupled with relaxed surface scan! All in a couple of simple #orca input and compound files! @faccts.bsky.social #compchem

150

Nicolas @nicolasn.bsky.social · 04/09/2026

I've been playing around with #Orca, setting up a compound script to run a constrained MD run on each step of a relaxed surface scan. From the MD I got a mean energy and standard deviation for each step in the scan. This allows IMO a better description of the dynamics of the process. #compchem

160

Reposted by Nicolas

Geoff Hutchison @geoffhutchison.net · 29/08/2026

A few weeks ago, I posted for the 2026 Avogadro Community Feedback Survey. Over 50 people responded, and several wrote multi-paragraph responses! I posted a summary to the forum and I want to include a few notes for #compchem discuss.avogadro.cc/t/avogadro-2...

discuss.avogadro.cc

Avogadro 2026 Survey Results

A few weeks ago, I posted the 2026 Community Feedback Survey - and I’d like to start doing this annually. Thank you to the 49 50 of you who filled out the development survey. Several people wrote…

172

Nicolas @nicolasn.bsky.social · 23/08/2026

New video showing my attempts at fixing the polymer builder tool in Julia, to make a chain of polyNIPAM for molecular dynamics! #compchem #julia #orca #chemcraft #programming #chemsky youtu.be/29Wi-BhcUpY

youtu.be

Trying to Fix the Polymer Builder: What is the best approach?

YouTube video by Prof Nicolas

030

Nicolas @nicolasn.bsky.social · 22/08/2026

I just want an Orcs for MD, where I just input my xyz, select the forcefield using a keyword, and it just works!! I can do this with XTB which is also awesome, but I have yet to test the limits on large systems.

000

Nicolas @nicolasn.bsky.social · 22/08/2026

Tried to install ambertools yesterday on the cluster. I needed the admins help to resolve colds versions, and then when actually trying to install it took forever and failed because of some system incompatibility. Why does any kind of MD have such a high barrier to entry?? 1/2

110

Nicolas @nicolasn.bsky.social · 18/08/2026

Yes I got really annoyed one time I suggested minor revisions but had a question for the authors I was interested in. I never got back the response and the published version didn't clarify that point enough to know what the authors thought about that issue.

110

Reposted by Nicolas

Thiago Carvalho @cyrilpedia.bsky.social · 18/08/2026

'But in April, CONICET announced the results would not be released until August 2027 because of a backlog of applications and cuts in the number of positions available. When the postdocs’ fellowships ended on 31 July, they were left without salaries and scrambling to find work.'

science.org

Budget cuts and delays leave hundreds of Argentine postdocs in limbo

Many scientists are scrambling to make ends meet and reconsidering their future in research

032

Nicolas @nicolasn.bsky.social · 11/08/2026

It was an honor and a pleasure to collaborate, focusing on the diradicaloid properties of these interesting complexes!

010

Nicolas @nicolasn.bsky.social · 04/08/2026

Short video showing #julia tools to parse Orca output files and get Löwdin charges and spins just from the optimized geometries in a relaxed surface scan. #compchem #chemsky youtu.be/E3MnlHahj84

youtu.be

Molecular Dynamics of Polymers with Julia: parse Orca output for charges

YouTube video by Prof Nicolas

030

Nicolas @nicolasn.bsky.social · 03/08/2026

These are awesome questions! Also related to beginning of life, self replicating molecules, etc. All fascinating topics.

010

Nicolas @nicolasn.bsky.social · 02/08/2026

I'm glad my "proposed method or framework" and "specific technical challenge addressed" were of interest to you!

Screen capture of spam email inviting me to submit a paper. The placeholder text where examples of my research should be have been left as in the template e-mail.

100

Nicolas @nicolasn.bsky.social · 01/08/2026

My quest to make an MD package for polymers is taking me to unsuspected places. I am now doing DFT relaxed surface scans on alkenes and biphenyls with various substituents to try to find general ways of predicting torsion parameters. Next it will be alkanes and dienes with uncommon substituents

040

Nicolas @nicolasn.bsky.social · 01/08/2026

I programmed many years in MATLAB and then Octave. Octave was too slow for some things, and I also tried python and it was even slower. So at some point I started learning Julia, and now I'm starting new projects in Julia. It has some learning curve but it's much faster

030

Nicolas @nicolasn.bsky.social · 31/07/2026

In case this reaches new people, reposting my first video on MD of Polymers with Julia series, showing a project I'm working on to generate polymer structures and do MD with them in a more or less automatic way. #julia #compchem #compchemsky #science #chemsky #academicchatter youtu.be/OrYavnmyyOo

youtu.be

Molecular Dynamics of Polymers with Julia - Intro

YouTube video by Prof Nicolas

183

Nicolas @nicolasn.bsky.social · 28/07/2026

A few flowers from my garden. Spring is close! The peach tree looks really good for this year!

000