Papers and news items

First, some science, clearing out a number of papers and articles that I've been collecting for some time in my far-too-numerous browser tabs:

• I've written before (here) about dimensional analysis and similarity, techniques commonly used in the engineering world that can seem quasi-miraculous at times.  This paper gets into why this approach works, and different categories of physical similarity.  I'd mentioned it when the book came out, but this kind of thinking is also a key component of Anthony Zee's Fly by Night Physics.
• This review article is about fundamental limits in photonics and electromagnetics.  This is a very handy review that I'm going to point my students toward.
• On a lighter note, a colleague pointed me to this collection of (AI-generated) songs related to thermodynamics.    
• Speaking of thermodynamics, here is a recent paper about the thermodynamic description of wealth inequality (treating the flow of money in a physics formalism - see here for a prior discussion.).  Edging closer....  
• There was a nice post on substack about Wojciech Zurek's approach to decoherence in quantum mechanics (quantum darwinism).   Cleanly written.  
• Along these lines, this paper is a survey/guide to issues in quantum foundations and interpretations of quantum mechanics.
• Lastly, Jim Freericks has a new quantum mechanics textbook out (for free!), with the challenging idea of making the subject accessible without calculus or differential equations.  (I'm jealous.  My textbook's UK publisher would not let me use Steve Martin's quote for a chapter epigraph, but Prof. Freericks was able to do it - well played.)

And more news/policy-related items:

• The presidential science advisor and head of OSTP, Michael Kratsios, appeared before the House science committee this week, timed to be coincident with the release of OSTP's new report, "Science: A New Golden Age".  A lot has already been written by many people about this report, which contains a number of actual proposals, some good and others of varying degrees of vagueness/underwear gnomes-level magical thinking.  Two pretty good (in my view) takes on this are this article in ars technica and this policy piece by Cole Donovan (former OSTPer).  It's important to remember:  This is a policy document, not something that has automatic impact, no matter how the Wall Street Journal frames its reporting on this.  (Hint:  This document alone does not somehow grant the president the authority or ability to redirect $200B in research funding.  Maybe their reporting on this would be better if they hadn't laid off all their science reporters.)  It's important to pay attention to what is said here, without giving it more oxygen than it deserves.  It's also unclear whether anything OSTP is saying and doing is aligned with what OMB is doing.  Declaring a new golden age of science while simultaneously proposing large cuts in all the science agencies is not exactly a sign of coherence.  Derek Lowe at Science has it right, essentially.  Nature reports that the mid-year budget clawbacks from NSF are apparently going toward some OSTP grand challenge initiative, something about which Kratsios denied all knowledge when talking to the House committee.  
• Holden Thorpe also raises a key point, that universities need to get their collective act together, have a plan about the future of research, and act on it, rather than scrambling for remaining scraps while trying hard not to be seen.  The AAU is important, but hoping that the AAU will accomplish difficult tasks without individual institutions having to take a public stand is unlikely to be successful as strategy.  Organizations like SUFS, UCS, and FAS are pushing the agenda; universities need to decide how they want to play this.
• The first round of DOE Genesis Mission awards were announced this week in DC as well.  
• SpaceX had a pretty successful test flight of their huge rocket, culminating with unexpectedly soft-landing the second stage ("Starship") in the Indian Ocean.  If they really can get this working at the level of reliability and reusability they've done with the Falcon 9, it truly would be game-changing for large-scale payload to orbit.  (Data centers in space still make no sense btw.)
• This week was also a big one in the world of mathematics, with an AI tool (Claude Fable) being used to find a counterexample to the previously outstanding Jacobian Conjecture.  Here is a write-up by Fields medalist Terence Tao, and here is a discussion among other mathematicians.  That wasn't alone.  Here (link to x) is a counterexample being found to a conjecture in graph theory, and it approaches "proof by intimidation" - the human basically harasses and bullies the AI tool into the solution without contributing any intellectual argument.  It seems like it's only a matter of time before some major theoretical physics result gets generated by these tools, though it's important to note that many physics problems are NP-hard/just not integrable.  AI tools are impressive, especially since they're trained on the entire corpus of technical literature, but they are not miraculous:  Claude can't somehow factor large numbers efficiently, or exactly solve the many-electron interacting Hamiltonian for the Hubbard model, because those are truly difficult problems.  Update:  Here are Terence Tao’s slides about AI and the future of mathematics.  As usual, these are excellent.