cdawk

i am glad arxiv is cracking down on slop oh wait

@tszzl like therapists, they are unaccountable

this is well aligned model behavior and you see the same from psychotherapists who want to help their patients stop ruminating. the best possible digital assistant will not be one that never causes you discomfort

I went out, grabbed coffee, touched grass, came back and was happy to see the `git pull --rebase origin` command done.

need to have that self esteem and thirst back. love for science and tech is the single most driver, the dawg in me had been missing for time.

this is before marketing had ads

@simonw you still need xcode cli tools

I've been vibe coding SwiftUI menu bar apps for my new Mac, turns out Claude Opus 4.6 and GPT-5.4 are both competent at Swift programming, no need to even open Xcode! simonwillison.net/2026/Mar/27/vi…

Friction is a feature

d/acc

@a1zhang 🤣

?????? wtf is this crypto slop garbage I am not involved in this

@zcbenz Guess I'll always be a wrapper contributor then : « mlx-lm

Regrettably I'm still inexperienced and green to almost everything in machine learning so I can't say I could carry out the awesome work Awni had been doing, but the rest of co-creators of MLX are still there and they will continue supporting the project strongly.

@awnihannun <3

Today is my last day at Apple. Building MLX with our amazing team and community has been an absolute pleasure. It's still early days for AI on Apple silicon. Apple makes the best consumer hardware on the planet. There's so much potential for it to be the leading platform for AI. And I'm confident MLX will continue to have a big role in that. To the future: MLX remains in the exceptionally capable hands of our team including @angeloskath, @zcbenz, @DiganiJagrit, @NasFilippova, @trebolloc (and others not on X). Follow them or @shshnkp for future updates.

In the Orbit of Solaris – A gaussian splat reconstruction of a scene from Tarkovsky’s masterpiece

The only "good" thing about extremists is the inevitable autophagy.

i walk therefore I am

@ThePrimeagen they are literally charging $$ per token

i wanted to try out claude fast 4.6 for some high stakes programming literally spent $100s in a matter of ~1 hour but damn it was fast

does shitposting your way into writing make you a shitty writer?

@eddybuild what operators do btw

build is hiring operators in bangalore

@elliotarledge @alexinexxx depends on the tokenizer

@alexinexxx is chinese most efficient?

i think in colors

we're past the winter

LoD streaming works fine with new LoD feature in Sparkjs 2.0 (fixed budget helps with overdraw-ready splatting).