half a week at the beach
Note: This post was imported and originally posted on leaflet on 2025-07-06
work
I was only working for half of the week before I went on vacation. Most of that time was spent on preparing my part of a presentation for DSO too. I also feel like I made good progress on some of my projects. One of the papers I've been on for a while had felt like a bit of a slog, but seeing the progress we'd made on it, got me more excited to work on it!
Looking forward to next week, when I can start getting used to the office and also get through my DSO presentation. There are some edits I need to make to my report too, but the deadlines are approaching, which means I'll have to get it done and then it'll be past me.
non-work
My partner and I went to Miami to vacation! It was really great for us to get some time away from work, especially with how busy she is at her job. There were scattered thunderstorms every day, but we still managed to get in a lot of sun, swim in the ocean, and see an Art Deco tour. It was definitely needed, and it was the first time I'd been to Miami Beach. Interesting, the layout of South Beach reminded me a lot of Hoboken! There's a Washington Street that runs through the grid in both, they have water to the East, and they're walkable (relative to the surrounding city). I still am very happy living and working in Hoboken, though.
As far as books go, I didn't read much on vacation. And my Oryx and Crake loan ran out when I was half-way through. While I'm on the waiting list for it again, I'm going to try to finish Talking to My Daughter About the Economy, which seems like a nice light book even though it's non-fiction.
things i saw this week
A ACLU push to avoid a 10-year ban on state AI regulations
Scott Aaronson posted about a new lower bound for Busy Beaver 6
questions
Do private constrained PRFs require structured assumptions?
This paper came out this week that showed that PCPRFs can exist and secret-key agreement not
By structured assumptions, I mean have specific complexity-theoretic implications, like hard problems in NP intersect coNP
What are the best attacks against LWE and LPN in different regimes?
This paper sparked my curiosity, and I realized that I know a lot of the generic-group attacks but not the lattice ones
I've seen using LLL to approximate SVP/CVP, but is that the best attack on LWE?
Also, LPN is such a weird assumption that I don't even know if the attacks would be similar.
Thanks for reading to the end! Here's a picture of a metal cat from Ancient Egypt that was meant to hold the mummified remains of a cat:
