Paul Boardman

everything 👏 is 👏 monocausal 👏 and 👏 specifically 👏 results 👏 from 👏 whatever 👏 shit 👏 I'm 👏 on 👏 about 👏 at 👏 any 👏 given 👏 time

I needed information from a dozen emails in my inbox... so I ran a screen capture tool, clicked through each of them and got Gemini 1.5 Flash multi-modal LLM to extract (correct, I checked it) JSON data from that 35 second video Total cost: $0.00082635 simonwillison.net/2024/Oct/17/vi…

Just plugged an iPad Pro into a 49" monitor to try/test @CellWalkApp, haha. This is so cool. This little rendering engine has come so far... 300 million atoms on a 49" monitor using a mobile GPU is crazy. #gamedev #scicomm #ios

Anyone interested in starting or ramping up in generative chemistry should read this paper and share it with their friends, students, and unrelated medicinal and computational chemists to help them get into the field. This is probably the most concise and useful review of generative AI in small molecule chemistry, which takes 20-30 min for experts to read but provides very deep insights (11 pages, mostly figures). If you want to see what published generative models and platforms worked experimentally for which targets and with what hit rates, you can print it out and get a cheat sheet. Of course, the most commercially tractable examples are rarely published but from what I can see - pretty much every published example was captured and scrutinized. This group reconstructed the timeline and did a gargantuan amount of work that an army of modern LLM agents would not be able to complete with a high degree of accuracy since many of the numbers are in image format and require additional processing. I was planning to write a review like this one day, but this group beat me to this and did a way better job as I saw some of the models I was not aware of. nature.com/articles/s4225…

Who dresses better? Alpha males or aristocratic babies? Let's explore. 🧵

We're excited to announce #AlphaFold 3 with @GoogleDeepMind in @Nature: our new AI model for predicting biomolecule structures with unprecedented breadth and accuracy. Expanding beyond proteins to tackle DNA, RNA, small molecules to fuel advances in biology & drug design 🧵

“The chief trick to making good mistakes is not to hide them – especially not from yourself. Instead of turning away in denial when you make a mistake, you should become a connoisseur of your own mistakes, turning them over in your mind as if they were works of art, which in a way they are.” — Daniel C. Dennett (1942-2024)

if you haven't seen this its utterly bananas the scale of potential pwnage seems almost unfathomable to me idk how you even start going about fixing the vulnerability that led to this

Phi29 phage "packages its 6.6-micron long, double-stranded DNA into a 42 x 54 nm capsid," requiring that DNA be folded to "near-crystalline density." The packaging motor exerts a force up to 57 piconewtons, about 20x more than myosin, the protein involved in muscle contraction.

@pauldambra More!

my very favourite meme format

@curiouswavefn Heisenberg certainly thought philosophy was relevant to physics. I really enjoyed his book "physics and philosophy". Possibly recommended by you at some point.

There’s been a lot of acrimonious debate recently about whether philosophy is relevant to physics. It’s why I find this paragraph from Eddington’s “Philosophy of Physical Science” so edifying:

🪩The @stateofai 2023 is now here. Our 6th installment is one of the most exciting years I can remember. The #stateofai report covers everything you *need* to know, covering research, industry, safety and politics. There’s lots in there, so here’s my director’s cut 🧵

wtf apparently the new superconductor works but only IN MICE

What a drama Lee and Kim has been disciples Korea University Dept of Chemistry head TS Chair. TS Chair comes up with a one-dimensional room temperature superconductor (SC) theory and publishes in 1994. Lee, the theoretician, takes up the mantle, publishes his Masters thesis on “Explanation of Superconductivity by the ISB Theory” in 1995. Meets Kim 1996. Kim is the experimental chemist and synthesis specialist. Lee and Kim discover a trace of the SC material in 1999. Lee finished his PhD in 2004, after a 9 year slog with the thesis “Theoretical Proposal and Synthesis of New Polymer Superconductors”. Kim graduates with his PhD in the same year (maybe same PhD shared between the two?) Both need to make money. Kim joins a battery materials company. Lee becomes a adjunct professor. Neither publishes at all. 4 years later, in 2008, Lee sets up Qcenter, which works on superconductivity while also doing consulting work for the big chaebols. Kim splits his time between the battery materials firm and Qcenter. Somewhere along the way, TS Chair passes, and his dying wish is that the students find the room temperature superconductor and prove his theory. In 2017, they start work on the final ramp. They have enough results to raise money. They go scrounge for dollars. I don’t think they get funded well. They get funding from some SME type industrial businesses, one in skincare and the other in auto parts. YH Kwon, a physics professor, joins in early 2018 joins the team as CTO but continues to hold his university appointment. By 2019 they’ve confirmed it’s a room temp SC and file a Korean patent for “production of low resistance ceramic compound”. 2020 Covid hits, and the team goes into monk mode. They finally isolate the SC, obtain its crystal and analyze its structure. They make a Nature submission and get rejected because of the Ranga Dias controversy. They’re asked to publish locally and get it peer reviewed. They bring onboard HT Kim from the US at this point (guessing on timeline). 2021 - they print another patent => mfg superconducting material at normal temp and pressures. They start work on chemical vapor deposition and are successful. They file an additional patent CVD is how you add this to our current manufacturing stacks. YK Kwon takes a sabbatical and goes off to Seoul National University at in Sept. Was he still part of the team after this ? Kim later posts an anti-physicist rant on LinkedIn. March 2023 - international patent filed. At the same time, Ranga Dias story comes out. JH Kim posts a comment to the story on LinkedIn “A retracted paper and an article in controversy immediately upon publication. It’s bittersweet. I really hope it’s findings are true”. In the 6 author paper, they very carefully lay out which of the authors did what planning, coding etc and what share of credit they deserve (equal, lead, support etc) It’s replicate or bust. They know no one will believe them otherwise.

Many of the drugs discovered by Paul Janssen and his team at Janssen Pharmaceutica have a piperidine ring at the centre. The reason why is a good lesson about how effective innovation happens in practice🧵

🐁 One of the most famous psychology experiments of all time, lost for 85 years because it was never formally published—because scientific publishing standards were already snafued—finally rediscovered! Many lessons here. gwern.net/maze

The rotating globe of the Globus INK (1967) showed Soviet cosmonauts their location in orbit. This electromechanical analog computer uses tiny gears for all its calculations, but it has an electronics board with relays to control the motors. Let's look at the electronics board.🧵

6th Artificial Intelligence in Chemistry Symposium lnkd.in/eTTBmbPS Now open for poster/oral abstract submission. #artificialintelligence #chemistry #rsc #AIChem23