2026-01-17 12:01:29
Went to see Marty Supreme. Wild, much different and better than I was expecting. Loved it. 🍿
2026-01-17 07:20:48
Fascinating drama in Elon Musk’s court filing against OpenAI. OpenAI’s web response is well done, with color highlighting of passages from Greg Brockman’s journal.
Always good to take notes. I wouldn’t be able to remember these kind of details from 8 years ago without writing them down.
2026-01-17 05:36:47
Rewrote some WebSocket code this week and it seems to have made a dramatic improvement not just for Nostr (which uses it) but general stability with hung Sidekiq processes. Reviewing Nostr, I still find it an elegant protocol. Just not sure how it fits in the modern social web.
2026-01-17 04:53:24
Ads are coming to ChatGPT. OpenAI is attempting to head off some of the inherent misalignment with ads and users by outlining some ad principles:
Ads do not influence the answers ChatGPT gives you. […] We keep your conversations with ChatGPT private from advertisers, and we never sell your data to advertisers. […] We do not optimize for time spent in ChatGPT.
The problem is once you’re making money with ads, it’s hard not to optimize for more ads. I just hope ads will remain a small part of their business compared to Plus and Pro subscriptions.
2026-01-17 04:10:20
This announcement from Wikipedia seemed to spark some debate. I think it’s only good news. If a company is significantly benefiting from Wikipedia and is able to help offset hosting costs, they should.
From the announcement:
Tech companies that rely on Wikipedia content must use it responsibly and help sustain Wikipedia for the future. One key way to do this is through Wikimedia Enterprise. Developed by the Wikimedia Foundation, Wikimedia Enterprise is a commercial product for large-scale reusers and distributors of content from Wikimedia projects.
It’s a little like supporting open source. Micro.blog has a recurring donation to the Hugo project. It’s honestly not very much money, but it’s a good habit and can increase as Micro.blog grows.
Molly White in a longer thread on Bluesky:
AI companies have ALWAYS been training their models on Wikipedia content, which under the free and open access model is available to anyone — including AI companies. Agreements like these require AI companies to limit and offset the strain they place on Wikimedia infrastructure.
AI using Wikipedia reminds me of the FAQ on setting up a Little Free Library:
I think someone is stealing books from my library and selling them, what do I do?
Remember that the purpose of a Little Free Library is to share books—you can’t really steal from it.
Wikipedia is an incredible resource. Even though I use AI chatbots all the time, I still find myself reading Wikipedia pages fairly often. There’s always going to be value in reading something longer that was researched and written by humans.
2026-01-17 01:11:24
I blogged briefly this morning about the W3C. Connected Places has a much longer post today about the new working group and participation.
I’ve submitted my name as an invited expert. I think they will have plenty of ActivityPub experts, but I’d like to help with testing LOLA and updating Micropub.