The authors show that, with the right algorithmic enhancements, first-order methods(*) can be competitive with state-of-the-art Linear Programming solvers, even for problems requiring high-accuracy solutions. This opens up new possibilities and computational trade-offs when solving LP problems, especially for large-scale instances.

• In numerical analysis, methods that have at most linear local error are called first order methods.

Blog: https://research.google/blog/scaling-up-linear-programming-with-pdlp/

Paper (go read the blog for intro): https://www.openread.academy/en/paper/reading?corpusId=235376806

https://dept.math.lsa.umich.edu/~idolga/physicsbook.pdf

What is Entropy?
John C. Baez, 2024

https://arxiv.org/pdf/2409.09232

What Goes Around Comes Around... And Around...
M. Stonebraker, A. Pavlo

It took three years to finish, but our follow-up to the 2006 "What Goes Around Comes Around" is finally out! Stonebraker and I examine the last 20 years in databases and discuss why relational databases + SQL will continue to remain on top.

https://db.cs.cmu.edu/papers/2024/whatgoesaround-sigmodrec2024.pdf

Control structures
Xavier Leroy
English translation, 2023-2024

https://xavierleroy.org/CdF/2023-2024/index.html

Merklizing the key/value store for fun and profit

https://docs.canvas.xyz/blog/2023-05-04-merklizing-the-key-value-store.html

docs.canvas.xyz

Merklizing the key/value store for fun and profit | Canvas Computing

A peer-to-peer runtime for decentralized applications