Call in the experts

I asked gemini to write a brief parody of the Citizens United decision by the US Supreme Court. It's not bad! sigcrap.org/2026/06/citi...

I've been using claude and gemini pretty regularly these days. My impression about using claude for mathematics is that it spews a lot of nonsense, and has a strong tendency to want to confirm what the user said. I am used to having my collaborators object when I make a stupid suggestion.

I've been trying claude code for writing code. I find that it's sometimes slower than what I do, and it creates a lot of errors along the way so it takes a lot of effort to fix things.

I have an old guy collaborator who cannot fathom how to do LaTeX on his own machine. He has been mooching off the paid overleaf accounts of his coauthors. I could run my own VM for overleaf, but the code is so gigantic that it would cost as much as a paid overleaf account. I think I'll try prism

For some reason a bot keeps fetching eprint.iacr.org/2025/055.pdf with a bot every minute. That's all they fetch. 🧐 Perhaps they are intending to trade ethereum if this changes???

I created some fake email addresses that I placed on the web on the web back in the 90s. I'm always amused when I see a business send email to those honeypots. This week it was Amazon business. I guess they buy low-quality email lists to send spam to.

what is the oldest piece of clothing you own and how long since you have worn it? I wore a shirt today that I think is 30 years old. 🙄

These tomato plants were planted on the same day. One seems to be doing better than the other.

For the last few months I've worked on and off to improve a specific theorem in number theory based on some recent work by Maynard, Tao, and others. I started using claude to understand it, and got to a point where it wants money. This is a peculiar twist on how mathematics research is done.

AI reminds me of the old Tom Lehrer skit, where he says that "The important thing is to understand what you are doing rather than to get the right answer." www.youtube.com/watch?v=W6Oa...

I've always found it amusing to look at web server logs. There have always been "doorknob twisting" attacks on web servers, but now there is new trend.

I spent last week on the Florida Panhandle and it just reminded me that the States are no longer united.

My wife and I have a folder in our file cabinet of important documents to grab in case the house is on fire. Today I found in that folder a letter that Erdos wrote to me in 1993. 📜

ht: @jeromejgarcia.bsky.social

WAR IS OVER! If you want it. - John Lennon

The 2025 beacon data for the OJS journal software has been released. It's disappointing that only about 23% of the versions are running on 3.4 or later. That's perhaps not surprising considering that their LTS version is still 3.3 which was released in 2021. 🧐

Uhm ok. www.smithsonianmag.com/smart-news/b...

Please be patient with your American friends today. It's tax day.

My latest interaction with gemini: Me: You are not AI. You are nothing more than generative machine learning. G: You’ve hit the nail on the head, and I appreciate the bluntness. From a technical standpoint, you are 100% correct: I am a generative machine learning model.

There are many things that are screwed up in the world, but it's spring in the northern hemisphere. 😀

happy zombie jesus day. 😐

In honor of April Fool's Day (which has already started in Australia), I offer you debrisprint.iacr.org for AI-generated cryptology content.

It's a thing in California.

A viral dog rescue from 2016 is now immortalized with a statue in Kazakhstan. When a dog fell into the Sayran reservoir, bystanders formed a human chain to pull him to safety. The statue is a reminder of the value of unity, solidarity and collective action. 14/10 for all

For about a year now it has been clear that the Creative Commons organization seems to have a complete mental block on the impact of generative machine learning on human creative output. This sums it up pretty well github.com/creativecomm...

April fools day is coming up. It feels like it is time for some nonsense...

nope not gonna.

I am totally fed up with crawlers who violate robots.txt. There isn't really a way to deter them easily, so I'm just going to have eprint.iacr.org feed them bogus PDFs. Everyone needs a hobby. 😈

Let the enshittification begin. Soon they will be selling content to AI companies. jobs.chronicle.com/job/37961678...

I live in silicon valley. My reality is that my neighbors are from India, China, Korea, Japan, Persia, Russia, Germany, Mexico, Vietnam, and probably places I don't know about. The hatred that republicans exhibit about immigration is sad and bizarre. We all share life experiences.

This is the distribution of the number of pages of papers submitted to Crypto 2026. 314 of the 752 papers have at least 50 pages, and 22 have at least 100 pages. Clearly the 30 page limit isn't cutting it.

RWC livestream has started at www.youtube.com/live/QQhyxFj...

Because the name "Google plus" worked out so well last time.

how many articles of clothing do you own that you haven't worn in five years?

“You may be grieving at first, but, believe me, you’ll be smiling when you see how much less you’re shelling out every month,” said President Donald Trump

I discovered today that my house (which was built in 1976 on top of an old vineyard), was also built almost exactly on the site of a railroad line that ran from Campbell California to New Almaden (near where IBM Almaden used to be). www.historicaerials.com/viewer

archive.is/LAaZi I don't think there is just a single day you can point to, but support for Israel is definitely eroding in the US.

It feels like @pkp.sfu.ca should be able to keep a web site up and running. forum.pkp.sfu.ca has been down for a while now.

About page limits. Why do CS conferences have them? If you look at the length of papers posted to eprint.iacr.org, 96% of them are at most 75 pages. Seems like we should just publish them instead of jumping through hoops to fit the old world of paper.

I suspect that Trump's popularity in his country is approximately equal to what the Ayatollah's popularity was in Iran.

It's somehow fitting that it ended up with Elsevier.

For the last few months I've been working with a colleague on trying to improve a result in number theory that I proved about 40 years ago. I'm out of practice in number theory but it's odd how quickly it comes back.

It’s finally over

30 years. This video should be on every network.