I’m off to a Dagstuhl workshop for the week, on creativity, programming, and generative AI. I’m excited to about charting this space with a great group for the week! Apologies in advance for all the slower email replies and other delays. www.dagstuhl.de/seminars/sem...

Submissions x 3

I had a lovely visit today at Lewis and Clark College, back in my home town. It was great to experience its lovely community, talk about justice work in CS, and connect with students and faculty around the future of CS!

Oh hi Texas, fuck you again. Love to all the kids and adults harassed and attacked in the coming years. Maybe I’ll join you next year in my trip to Austin. Pee on the capitol, anyone? www.axios.com/local/housto...

Once again, fuck Texas, and love and solidarity to all the Texans resisting its bullshit. www.nytimes.com/2025/09/26/u...

In case you haven’t followed along, Erin’s got a nice catalog of recent escalations on trans rights and safety in the US. And I’m feeling it at home.

This city is at times the safest place to be, and at other times not safe at all. It is increasingly a destination for hate speech and hate crimes. Time to get strong, together. www.thestranger.com/news/2025/09...

They were always going to be bullshit machines; it’s in the architecture. The question continues to be in what contexts a bullshit machine is valuable, rather than destructive, and whether net, those contexts justify the destructive uses. That is ultimately a moral question, not an engineering one.

The current goal is to implement prior to world cup. We'll see if that actually happens, but at least there's a goal and pressure to reach it: seattletransitblog.com/2025/01/23/o...

Ok my cis friends, we are going to need a lot of you to step the fuck up and fight hard for us

www.them.us/story/trump-...

I think someone might have forgotten to send me my woke mob radical leftist trans terror cell membership card. Is there a meetup? Are there cookies? Can I bring my cat?

I’ve had so many challenges with that kind of data. Where to store it, who gets access, how to contextualize it if I have context for the cause, when to delete it. Collecting is easy but managing it thoughtfully is hard.

Good point. Need better training data.

"crackpot with a bullhorn" track is a great name for the internet and LLMs

PSA from your neighbor
youtu.be/JMuJcrolgUA?...

I think there’s a lot of potential in that too, at least as a signal for whoever is deciding whether to compensate. It’ll be fun to play these out for the journal and then share what we learn. I have a list of 80 editors across academia who want to be fast followers if it works.

The vision at the moment is that a program chair/associate editor would make the final call on whether a reviewer did their job. If it doesn't reach the community defined quality bar, it doesn't get compensated. But this would be up to communities to define and communicate.

Actually, we are thinking of making a tipping feature :) You like a review, tip a token, and maybe people start aiming for great reviews.

This is the experiment we're trying next: reciprocal.reviews. Earn tokens for reviews by writing adequate quality reviews, spend them on submissions. Earn the privilege to publish. We're at 70% on the beta, and should launch an ACM TOCE pilot next year some time.

So lame. And agreed on training; I've built a lot of these trainings for conferences, and the invested folks engage and follow them. But the unaccountable, uninvested ones don't. I think that goes back to the lack of accountability. And then folks like you get punished for their irresponsibility :(

That's super hard to do when there are too many papers, not enough reviewers, and no accountability for writing reviews within the bounds of a community's norms. Seems like we should be focusing on how to make more invested reviewers than how to map a shrinking set of invested people to each other.

Incentives for review quality are incredibly weak: near complete anonymity means no accountability, reviewing needs mean no leverage over review quality, and institutions elevate writing over reading. The only way to disrupt that is to change incentives. There are lots of ways to do that.

Lots of reasons to be skeptical about content like this (bothsidesism, platforming, and cringey terminology), but I like how it scaffolded perspective taking, organically centered love, and modeled openness to listening and learning. youtu.be/mQlarHFPoaM?...

iSchools are a great start! We've been training hybrid scholars of this kind for more than two decades. Every university should have one; more than 130 universities do already. As a computer scientist who stumbled into one, it's been transformative for me.