David Gleich

dgleich.bsky.social

14 October 2026 19:39

On LLMs in research/theory/math. I feel like the difference between how people value this is simple. If you find the value is in the outcome (the theorem), then LLMs are awesome. if you find the value is in the journey (the way to find the proof), then LLMs ruin it.

Badger Punch Games

badgerpunch.com

22 July 2026 19:32

Roguecraft DX is a Lovecraftian roguelike where even the dungeon has a soft side.😊 - Find caged creatures in the dark - Free them, pet them, earn their trust - Watch your new familiar fight for you, and even fetch you a potion Yes, you can pet the familiar. We insist. #WishlistWednesday #pixelart

James Andrew Smith

drsmith.bsky.social

15 July 2026 22:17

As The Mathworks doubles-down on including AI as default in #MATLAB, I'm seriously looking at alternatives for teaching procedural programming. Looks like Octave might have an Arduino toolkit... gnu-octave.github.io/octave-ardui...

Olivia Guest · Ολίβια Γκεστ

olivia.science

11 July 2026 12:13

We analyse this through 1️⃣ the cognitive: what is forfeit when statistical pattern-matching substitutes our thinking 2️⃣ the pedagogical: AI literacy as currently deployed is a symptom 3️⃣ the political: AI's infrastructure inherently naturalizes labor displacement pdf: zenodo.org/records/2122... 5/

Stephen McCarthy

misterjabsticks.eurosky.social

02 July 2026 07:00

Commodore claims that it uses AI "sparingly, responsibly, transparently, and only where it adds meaningful value without disenfranchising people." It is not possible to use GenAI responsibly.