I demand credit

I think it's preposterous to think that survey takers aren't also just asking chatgpt to pretend to be a sag writer or whatever and copy pasting some subset of answers.

I appreciate the (admittedly biased) attempt to measure this but I'd need to know more about how they distinguished. It doesn't have to be bots either; we have seen instances of 30% cheating in classes.

Seeing lots of papers at #CHI2026 still using online survey tools... not sure if we can still do that.

I also don't understand why this is a problem.

So can anyone tell me what cutting a submarine cable even does? There's redundant links so things will slow down a little but not a lot of impact on connectivity. Yet they do it. There must be some reason they are so geopolitically relevant.

Torsion is rough.

I'm not experimenting on that; I think it's common knowledge that mouse behavior is part of the model. Could be wrong though.

I once again remind everyone that "finding new customers for a product" is not the same as "one person one vote". No one is democratizing anything here.

Concur, though they still seem to work, which is interesting.

Yeah but are you saying they don't work? I think they do. There are a few specific solvers but nothing generic I can find.

And yes, you're going where my head is, though I'd really like real-world captchas (maybe for exams?)

Can anyone tell me how captchas still work? It seems like they shouldn't but I still see them around.

How Cultural Adaptation of LLMs Can Backfire The surprising insights that made me rethink the goal of pluralistic alignment

Working on pluralistic alignment? Our newest Substack post dives into Jeff Basoah’s research on whether and how AI should culturally adapt. globeai.substack.com/p/how-cultur...

Unveiling and Engaging with the Humans of Networking Research — Internet Society Pulse Action research, Participatory Action Research, and Participatory Design approaches should be employed in Internet research studies to integrate community participation directly into the research outp...

Nova made a nice writeup of our recent hotnets paper on human methods in Networking for the ISOC blog: pulse.internetsociety.org/en/blog/2026...

NINeS 2026 is tomorrow! We have just posted (at nines-conference.org/attend) details on how to join the livestream, which starts at 1am UTC, and will let you experience what is going on at several pods and also hear our keynotes. We hope you join us.

1/2

I am not sure what to take from that.

With the recent news of the pullback, I am troubled a bit that it seems that Alex Pretti simply bringing his gun, not using, pulling, or brandishing, was really impactful in this change. If he had been killed without it, the uproar seemingly would have been a lot smaller.

Like, sure every system is developed within a political environment but how much is there and how intentional was that are really important things to think about. It's ok to strive to build "apolitical" things even if you can't get there in reality. It just ends up meaning "less political".

I find myself increasingly frustrated with the "artifacts have politics" line of reasoning; it's correct and good to understand (we now teach it in our intro) but the degree and intention of the politics is way more important than the binary nature of it.

Really delightful book fwiw.

Allen School researchers explore smart technologies for empowering people experiencing homelessness - Allen School News Modern Internet of Things (IoT) technologies such as smart home devices, sensors and security cameras have transformed how we interact with our physical environments. As University of Washington resea...

Can smart technologies help people experiencing homelessness? #UWAllen prof @kurti.sh and postdoc Esther Jang worked with tiny house villages in Seattle to explore their potential for managing infrastructure and other community needs. #ThisIsUW #CSCW #HCI #IoT news.cs.washington.edu/2025/12/04/a...

I don't think it's well understood that Google lost the SEO wars. LLMs/AI let them fight back by controlling the content but there's no going back to a world where Google can correctly identify good human content. It's just not possible.

Jon E. Froehlich Professor Froehlich’s academic webpage.

I am recruiting new PhD students this year in:

🌆 Urban Computing / GeoAI
🥽 Augmented Reality
🎨 Creativity Support Tools

Why join us? Work on state-of-the-art Human-AI research that has real-world impact at one of the top CS programs in the world.

Deadline: Dec 15
See: jonfroehlich.github.io

I once again implore technologists to stop saying they "democratize" things. You're not doing "one person one vote", you're just selling to new people that didn't get sold to before.

Is the scam the job, the human trafficking, or both?

I'm going to assume this is less ominous if you can speak the language.

Higher than ever is a bit of an overstatement, 2020 does seem to have been the peak. It's still well above the norm for the last few decades.

I don't think this accurately matches the dynamic; overall political engagement (I'll measure it as voters in the presidential elections, though I imagine there are others) is higher than ever. This doesn't always mean that more people are arguing over kitchen tables, but that's my experience too.

Gotta find someone to play with right?