TechTakes

The myth that talking about bubbles prevents bubbles for when someone starts saying some “oh but everyone says AI is a bubble which must mean its not”

Hey, remember Pascal’s wager? What if we got the leader of the church to be part of our new “ai ethics” product launch https://www.bloomberg.com/news/articles/2026-05-18/anthropic-s-co-founder-to-launch-encyclical-on-ai-with-pope-leo?accessToken=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzb3VyY2UiOiJTdWJzY3JpYmVyR2lmdGVkQXJ0aWNsZSIsImlhdCI6MTc3OTExMTY3NCwiZXhwIjoxNzc5NzE2NDc0LCJhcnRpY2xlSWQiOiJURjhGMjNLSkg2VkcwMCIsImJjb25uZWN0SWQiOiI4QjVENTdBQTJGNjY0RTdCOEQ4NkUxMzcwMDNDMkVDNCJ9.OEFYqROFrqezWHTRznaz9xKZCem4PIY99uXzZJ3bWQQ&leadSource=uverify+wall

Heise was gushing about (1) Google & OpenAI working together on tagging AI output and (2) camera-makers trying to push cryptographical attestations for real photographs.

Felt the need to point out that (1) is just fear of Habsburg-AI and (2) is a road to a DRM-like unworkable dystopia.

new tante https://tante.cc/2026/05/20/on-google-declaring-war-on-the-web/

GitHub got pwnd by a compromised VS Code extension. Internal repos were exfiltrated.

Disclosure on Twitter

Elon Musk has lost his court case against OpenAI

Google released their new Gemini 3.5 "flash" model at I/O yesterday. For those who aren't familiar, the "flash" model is typically marketed as the lower end and the "pro" model is the higher end for each given model generation.

The interesting thing here is that the new "flash" model is almost as expensive as the "pro" from the previous generation.

As my favourite "neutral-but-not-really" AI booster Simon Willison says:

This fits a trend: OpenAI's GPT-5.5 was 2x the price of GPT-5.4, and Claude Opus 4.7 is around 1.46x the price of 4.6 when you take the new tokenizer into account.

It feels like all three of the major AI labs are starting to probe the price tolerance of their API customers.

Speed running enshittification - a process that typically only works when people are reliant on your product and have no other option than to pay the inflated price

HN, lobste.rs and LW are jizzing themselves over the ~~news~~press release that ChatGPT[1] has disproven an Erdos[2] conjecture.

no-one pauses to think this is a transparent attempt to goose interest in OpenAI before they commit an IPO

[1] I know it's not literally ChatGPT, it's an "internal model"

[2] fuck trying to find the double acute accent or w/e it's called over the o

This from a while ago but I forgot about it until today: Eliezer jumpscare in this interview about Absolute Scarecrow (its very brief but is there)

on the topic of EY's book, the ratings on Goodreads have slowly crept down (from 3.97 to 3.92)

and on the topic of ratings, the AI Doc has also gone down to an 6.9 on IMDB

James Gleick:

This author—caught using AI to make up quotes for his book about the dangers of AI—has the gall to say it proves him right. You can't trust him, so you can't trust anyone.

Iran has created an Insurance Company to guarantee safe passage through the Strait of Hormuz. Vessels pay their Insurance Premiums in Bitcoin

https://x.com/IranObserver0/status/2055716544596922700 via naked capitalism

this is fine