"The complete and utter failure of the metaverse is a reminder not just of the fact that the future Silicon Valley is force feeding us is not inevitable, but that quite often these oligarchs quite simply cannot relate to real people, don’t know how or why people use their products, and very often have no idea what they’re doing." https://www.404media.co/rip-metaverse-an-80-billion-dumpster-fire-nobody-wanted/, Jason Koebler

"If we believe in the dignity of work, can we justify taxing labor at a higher rate than earnings from capital gains and dividends?" -- interesting discussion with Michael Sandel, https://www.noemamag.com/the-quest-for-contributive-justice/

Glitches in LLM training data are hard to remove, but they're useful to catch AI slop. And sometimes entertaining. https://theconversation.com/a-weird-phrase-is-plaguing-scientific-papers-and-we-traced-it-back-to-a-glitch-in-ai-training-data-254463

If you need a few podcasts to listen to over the break, and you feel passionately about saving the Internet, do check these out: https://www.eff.org/deeplinks/2025/10/effs-how-fix-internet-podcast-2025-review

Since prompt injection is an unsolved problem, agents exposed to web content are not trustable by default. To make a safe browser agent, we need layers of deterministic and model-based protections. I wrote about what we're building in Chrome: https://security.googleblog.com/2025/12/architecting-security-for-agentic.html