Kyle Howells

⭐️Spherium 1.0 Released! Designed to view and share 360 images from your Photo Library. AppStore: apps.apple.com/us/app/id15021…

GPT 5.4 in codex implemented TurboQuant in MLX in like 25 minutes by giving it model weights and the PDF report Sorta insane this is where we are now

You can't convince me otherwise.

The way that Apple doesn't care about all the little flashes within this small iOS interaction bothers me immensely Apple's software interactions used to be state of the art, the envy of all product designers Now they're so janky I can't bring myself to use them in my app

Codex laughs at your petty guardrails

some cool toolbar icons from iWork '09

@sids7 Connect to Hopper MCP and it’s basically as good as having access to all of Apple’s internal source code

Wasn't joking about this one btw You can reverse-engineer pretty much any part of Apple platform internals in seconds using Claude or Codex with Hopper MCP

Using agents to code review is also very useful. Normally based on what the code is doing you have an idea of certain pitfalls, or "oh, I hope they didn't implement it that way". So asking an agent "Does the PR use X technique?" is a quick first pass for any main worries.

I compared UIKit and SwiftUI performance with the most ridiculous scroll view possible. Memory usage wasn’t dramatic, but was interestingly spiky. Perhaps due to the different numbers of gifs and different resolution of the images rendered on each cell. When it came to profiling peak memory and energy usage, I expected UIKit to yield the same kind of results I saw in SwiftUI: We’re still displaying an enormous number of elements, animating them the same, and these images are not any bigger or smaller, right? Now don’t I look silly. Memory usage hovered around 92MB for UIKit, compared to 248MB in SwiftUI. Read my scientific performance comparison right here 🧪 blog.jacobstechtavern.com/p/swiftui-vs-u…

Passthrough access isn’t available to visionOS apps. So I built it. 😉 Introducing PeekABoo — the first consumer-facing passthrough API for visionOS.👀 It unlocks a whole new category of spatial apps. Here’s how I pulled it off 👇🏻

I've been working on an MLX port of Moonshine and think I've worked out a way to boost not just the speed, but the actual model accuracy over the official implementation. On various test files I've been using the 35m tiny model now matches the WER of the 250m medium model.

And now I've submitted the Swift version of DeepFilterNet MLX support. github.com/Blaizzy/mlx-au…

Submitted my first PR to the MLX-Audio project to add DeepFilterNet noise reduction ML library.

The git repo: github.com/moonshine-ai/m… The research papers: v1: arxiv.org/abs/2410.15608 v2: arxiv.org/abs/2602.12241

For context the first version of Moonshine took the Whisper model and optimised all the parts which made latency slow (the 30s fixed size audio window). V2 takes this further reducing the latency and size even more!

Moonshine is a really cool Speech to Text model optimised for super low latency on low resource devices like PI's and phones. About 100x less latency than Whisper. Spent some time this weekend working on adding support for word level timestamps to it github.com/moonshine-ai/m…

The MacBook Neo uses an A18 Chip that is in my iPhone Pro Max, and it runs full macOS competently in eight gigs of RAM I want to plug my iPhone into my thunderbolt dock and run macOS X. It doesn’t seem like it’s a technical problem anymore, now it’s organizational willpower

@Prince_Canuma Thanks for the tip 🙂

@Freerunnering Protip: Add your X account to your GH account profile so I can easily tag

@Prince_Canuma Thanks, got a few more coming in soon, including swift versions.

@Freerunnering Thank you for your awesome contribution🚀

🚀 mlx-audio v0.4.1 is out! New models: → Granite Speech 4.0 (STT and AST) → Canary STT (NVIDIA canary-1b-v2) → Moonshine STT → MMS STT → FireRedASR2-AED STT → SenseVoice STT → Fish Audio S2 Pro TTS Plus: → Native MLX DeepFilterNet speech enhancement (v1/v2/v3) → OGG, Opus & Vorbis audio format support → LID fix for ECAPA/SpeechBrain alignment Thank you all contributions for this release: @lllucas, @beshkenadze, @andimarafioti, mm65x, irachex and Kylehowells! 🚀 > uv pip install -U mlx-audio Leave us a star ⭐️ github.com/Blaizzy/mlx-au…