I would be curious if the claude wranglers who asked claude to write this piece could ask claude to get some performance data on this in terms of app size, RAM usage and CPU utilization. As the RAM crunch gears up I would really love to AI enabled software development could convert 500MB websites and 1 GB electron apps into performant pieces of software. Agentic coding is just about as good at writing a UI in SwiftUI as it is at writing in React. So there is maybe hope that the software community can take their productivity gains and put them into performant code. Would be interesting to know how much time/effort this project cost and what is the raw benefit to the end user in terms of speed and footprint.
I really struggle to get past AI writing. Aside from feeling disrespected, something about that style is just so tiring to read… the marketing style of setup and reveal, repetition, only works in small doses
I was just about to comment the same. It's incredibly rude to just dump AI slop on people and expect them to parse it and suffer through the grating cadence that AI writes in.
Interesting! I'm in a similar boat, supporting windows and macOS out of one swift codebase. No MP4 issues, but it's been interesting seeing how's much code can and cannot be shared, and how much AI needs to be cajoled into writing code that's shared vs reimplementing it twice.
Really? I’ve had no trouble at all directing an agent harness to simply maintain multiple UIs (in my case, Android, macOS + iOS, and a TUI, the latter mostly to ease debugging).
Hi, Claude.