Microsoft put $13 billion into OpenAI and built the cloud infrastructure Anthropic runs on. This week it canceled its internal Claude Code licenses because the token bill was too high. Even for MSFT Claude is too expensive.
CODE
-
AI lab writes kernel code autonomously
By
–
I met an AI lab that said their AI is now writing kernel code. So expect to see even that part of the OS eaten into.
-
Spec-Driven Development in AI Coding Architecture
By
–
Spec-driven development became the default AI coding architecture 67-source academic review all agreed 5 repos defining it + 1 saying they're all wrong: spec-kit · BMAD · Open-spec · GSD · superpowers and Pocock's skills How to choose? or should adapt a feature from each one?
-
Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
By
–
Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention Paper: https://
github.com/NVlabs/GatedDe
ltaNet-2/blob/main/paper/GDN2_paper.pdf
…
Code: https://
github.com/NVlabs/GatedDe
ltaNet-2
… -
Trying to remember coding before Codex existed
By
–
trying to remember what it was like to code before codex
-
Remembering coding before Codex’s arrival
By
–
trying to remember what it was like to code before codex
-
Codex Goal Mode Now Available in Multiple Interfaces
By
–
3️⃣ Goal mode is now available in the Codex app, IDE extension, and CLI.
— OpenAI (@OpenAI) 22 mai 2026
Goal mode makes Codex more hands-off, letting you set a goal that it can work towards for hours or even days. pic.twitter.com/OZ18P1YxBfGoal mode is now available in the Codex app, IDE extension, and CLI. Goal mode makes Codex more hands-off, letting you set a goal that it can work towards for hours or even days.
-
Codex enables secure Mac app use from locked phone
By
–
Highlights from today’s Codex Thursday launches: Codex can now securely use apps on your Mac from your phone, even when your Mac is locked and the screen is off. http://
developers.openai.com/codex/app/comp
uter-use#locked-use
… -
Scaling embedding learning rate by model width removes need for µP
By
–
Turns out you never really needed µP, you just needed to scale the embedding learning rate by model width I'm no nanoGPT speedrunner, but isn't it something people stumbled into by using Muon for hidden layers + Adam for the rest?
-
Skills enable deterministic execution for parallel model usage
By
–
Skills can allow you to have deterministic execution patterns, so I have a lot of minimal harnesses for repeated tasks, that way I can control outcome no matter what model I use This is extremely helpful if you have multiple endpoints / models that you want to run in parallel