work

This past week was mostly just a grind to get all of my in-person work in a place, so that I can work remote this coming week. For the most part, that wasn't too hard — just releasing and prepping things for class, rescheduling some meetings, and working on a few papers. As far as I can tell, everything is in a good place! I will still be working full time this coming week too, but from the west coast. The main targets now are to work on my FOCS presentation and my (primary) STOC submission. If all goes well, I'll also work on a second STOC submission, but even if that falls through, I'm sure we'll find a place to put the ideas.

The main thing slowing me down this week was a mild existential crisis brought on by absolutely breakthrough results released by Anthropic and OpenAI. There isn't much I can say that hasn't already been said. I'm mostly just in a state of uncertainty about the field and the future. There will clearly need to be a shift in how academics are judged and how they're used. I don't know what this means for me and my own journey yet. I hope, however, that the community (especially the younger members) can push for a new system that is workable, fair, and useful. I don't have a solution, but hopefully by exchanging ideas and some creativity, we can have really great institutions. (I see this also as an opportunity to fix other, longer standing, problems with academia too.)

All this is without even mentioning the potential threat to cryptography alone! None of the breakthrough results had to do with breaking cryptosystems, which honestly is a little suspicious to me. There is some fear that these results have been withheld, for the companies to follow best practices for responsible disclosure of vulnerabilities. I certainly hope this isn't the case, but either way, I expect a lot of disruption in both theoretical and applied cryptography as we potentially scramble for new parameters, assumptions, and standards. (Or maybe we'll be able to settle P vs NP and prove some systems secure unconditionally :-).)

non-work

The most exciting thing outside of work I did this week was fly to San Diego! I arrive on Friday afternoon and have been spending time with my partner, Michelle, out here (hence the later post). We had a lot of fun the past couple of days, going to our favorite brunch places, adventuring to new parts of the city, getting some pumpkins for fall and more! Even now, I'm writing the post from a nice pastry and coffee shop that Michelle found. It's been a very relaxing and pleasant long weekend already, and in the coming days, we'll continue to find fun things to do. I'm very appreciative that Michelle has done her research and knows about so many fun things to do and has made the time to do them with me.

I also finished the second book of Dungeon Crawler Carl on the plane and am onto book three! It remains just such a great series, and I expect to keep reading it. (Although, I also am hoping to read or listen to Fires of December soon too!) I'm still having a lot of fun with fiction, so I don't now when I'll switch back to non-fiction. There's still a bit of a queue of those books, so I'm sure I'll make the switch eventually.

questions

• What is a good resource for learning to read and write Lean 4 statements?
  • It seems like this is going to become a central skill that's required for future theory research.
• What is the role of academics as AI reaches a more mature state?
  • It seems like the pursuit of truth is cheap, if for example, we can run models which can formalize or find counter examples to theorems.
  • Is the role then to choose which theorems are most "intrinsically valuable?" Or, as a pragmatist, would we find the most "useful" theorems?
  • What if AI eventually becomes better at identifying the research directions which are "best" to pursue?

Thanks for reading to the end! Here's a picture of Michelle's cat, Georgino, and some squid game figurines we saw in a store.