Robert Palgrave

...but them being hard to make means they will have to be even better at their job to make manufacture worthwhile. A difficult conundrum to solve.

To me computation seems much better as a focused tool for understanding a material of interest, rather than a discovery method.

Fact is shovelling stuff into a furnace is pretty easy. And diffraction is pretty easy. That accounts for so many oxides and metal alloys being known. Typing to beat the furnace in discovering those - I'd say good luck.

Success will come from niche compounds that are hard to make

As far as I know, none of those kinds of materials can be discovered with current ab initio methods. But even the material types that are idealised for computation - crystalline, simple composition, no disorder... are there examples of these discoveries led by computation? I'd say not many.

I'd also note the original table was about materials discovery specifically, not calculation in general.

Materials can be non crystalline, or composite, or nanoscale, or have texture, impurities, or compositional disorder. In fact most useful materials fall into at least one of those.

to a high degree without computational theory (meaning an ab initio) Examples are all the proteins and enzymes Hodgkin solved, all the structures Braggs solved, all the organic molecules that were essentially solved before even diffraction (Kekule - even if some had to rely on dreams)

The table is tongue in cheek for sure... and of course we need both experiment and computational theory.

I would also say although there can be no observation without theory (theory in the most general sense) - in materials science it is quite possible to work out 'what this stuff is'...

A table comparing the computational power of HPC and a furnace

Pocket guide to materials discovery calculation methods (repost from the other place)

Add to cv anyway

Robert Palgrave lecturing on the role of AI in materials discovery

@robertpalgrave.bsky.social delivers the PCG plenary lecture at #BCA25 "The role of AI in materials discovery"

This is why everyone falls in love with solid state chemistry - first year lab straight out of Harry Potter

Nice visit to the European Commission to discuss AI in materials!

The UK National XPS Service now on bluesky!

You can select the style, it can be anything! Gospel, prog rock, nasheed... In any language too...you just enter the lyrics and describe the style.

The website isn't specifically for academic papers, but maybe we can create a new genre!

Original paper here

papers.ssrn.com/sol3/papers....

Lattice by @xpsbeats | Suno [Spoken, Synthwave, Trance, Heavy Trap Drums, Dubstep Breaks, Neon Vocals, Retro Futuristic, trance] song. Listen and make your own with Suno.

Hey it's 2025 so here's a version of our Lattice paper as a AI generated trance track

suno.com/song/a512663...

Yes very much agree, but I think electroboom had the better video on this one topic

YouTube video by ElectroBOOM How Wrong Is VERITASIUM? A Lamp and Power Line Story

Also see the rebuttal from the incomparable electroboom

youtu.be/iph500cPK28?...

YouTube video by Veritasium The Big Misconception About Electricity

Interesting video on this

youtu.be/bHIhgxav9LY?...

Facts are invented not discovered

Twitter is getting really really weird

x.com/jlippincott_...

It's always been clear most out-and-out racists voted Tory. I'd always wondered what would happen if their votes were subtracted. Last GE we found out.

If left is reasonably united Tories can't win without the Reform vote. That's why they are scrabbling to the right.

I do appreciate all your replies and definitely not tl;dr!

Happy Christmas!

These are the people I'm replying to. Their mental model is that all the IR is absorbed as pure CO2 does in a short tube because if its high attenuation. It *is* a superficially convicting argument.

By showing that some IR is escaping earth that proves that model is wrong.

Thanks for all these helpful posts.

I do think this that when laymen deniers use 'saturated' they mean it in a different sense. They mean all the IR is absorbed by CO2 and none is left to absorb, based on simple ideas about absorption coefficients.

E.g. these replies on twitter

I understand that CO2 emits as well as absorbs, and that the 650cm-1 light escaping earth has been absorbed and emitted many times

My point was to address the specific wrong claim very often used by deniers with as direct an observation as possible, which they'll find hard to dismiss

I thought the best observation to disprove this is to show that in fact 650cm-1 light is radiated from earth, showing that the atmosphere isn't saturated in the way described in those papers

What 'they' mean by saturated is that 1 atm air absorbs 650 cm-1 to 99.99....% over a small distance compared with the thickness of that atmosphere. They conclude that none of the IR from ground reaches space, so adding more CO2 has no effect.

The saturation theory is tested and falsified by these observations.

Even though it seemed intuitively right, it is in fact wrong. That's what makes it the ideal tool to keep pushing climate change denial.

Mlynczak et al. GEOPHYSICAL RESEARCH LETTERS, VOL. 33, L07704, doi:10.1029/2005GL025114, 2006

Another balloon flight from 2005, this time to 27km

Again CO2 not saturated.