distrust any business or enterprise that claims otherwise. They clearly don't understand the concepts of agency and accountability. And you don't want to deal with anyone who hasn't realized what makes them different from a machine. You'll be disappointed.

My fellow UX/UI designers, please welcome: Google's Material Design 3 Expressive

One of the people behind the recent incredible performance improvement of Typescript has been laid off by Microsoft. He was a Microsoft employee for 18 years.

Man, nobody told me Gemini Advanced on 2.5 Pro was this good. First drafts always look great, no need for revision. Its answers are just straight to the point, no introductory bootlicking like "oh yours is a very good question". It gulps down whatever context file and doesn't need any iteration or prompt fragmentation.

I love it. Definitely not going back to chat gippity after this.

Tsciao, currently streaming on Twitch (in Italian):

https://www.twitch.tv/callmezulo?sr=a

Yesterday Microsoft killed the paid AI code editor market (in a good way):

"We will open source the code in the GitHub Copilot Chat extension under the MIT license"

https://code.visualstudio.com/blogs/2025/05/19/openSourceAIEditor

I’ll attach the paper: https://comsec.ethz.ch/wp-content/files/bprc_sec25.pdf

Okay, this is a bit Cicero pro domo sua, but hear me out, okay? ;)

Read my latest article: In Defense of LLMs

We welcome a new admin in the channel! 🎉✨💯

Founder admin’s silly project:

https://github.com/atsetilam/word-by-word

Huge GitHub milestone. Repo #1B was created today and it's named “shit”.

I may be late to the game but this research paper, “Simulating Time in Square-Root Space” by Ryan Williams seems very interesting.

Abstract:

We show that for all functions t(n)n, every multitape Turing machine running in time t can be simulated in space only O(tlogt) . This is a substantial improvement over Hopcroft, Paul, and Valiant's simulation of time t in O(tlogt)  space from 50 years ago [FOCS 1975, JACM 1977]. Among other results, our simulation implies that bounded fan-in circuits of size s can be evaluated on any input in only spoly(logs)  space, and that there are explicit problems solvable in O(n) space which require n2− time on a multitape Turing machine for all 0, thereby making a little progress on the P versus PSPACE problem.

Our simulation reduces the problem of simulating time-bounded multitape Turing machines to a series of implicitly-defined Tree Evaluation instances with nice parameters, leveraging the remarkable space-efficient algorithm for Tree Evaluation recently found by Cook and Mertz [STOC 2024].

yoooooo we’ve got V-JEPA 2 before GTA6

GOOGLE THAT’S NOT WHAT I MEANT

What do you guys think of tea.xyz ?

After 20 years, PNG 3.0 is here. It brings HDR, APNG and Exif support for higher image quality and greater versatility.