Hacker Newsnew | past | comments | ask | show | jobs | submit | alwillis's commentslogin

> It seems clear to me that they use OS updates as a way to eventually slow your device down

This sounds like an exaggeration of what happens after an upgrade: iOS has to re-index your entire phone for Spotlight, etc. Same thing for Photos if there have been changes.

Depending on which phone and the amount of storage, your phone can feel kind of sluggish for a while until the background indexing is done.

If you update before you go to sleep, your device will be fine in the morning.


This should get you caught up: https://nerdy.dev/cascading-secret-sauce

Hopefully the next time it’s updated, it should ship with some variant of the M5.

> I’m expecting Apple to release a new Mac Pro in the next couple years

I think Apple is done with expansion slots, etc.

You'll likely see M5 Mac Studios fairly soon.


I’m not saying a Mac Pro with expansion slots, I’m saying a Mac Pro whose marketing angle is locally running AI models. A hungry market that would accept moderate performance and is already used to bloated price tags has to have them salivating.

I think the hold up here is whether TSMC can actually deliver the M5 Pro/Ultra and whether the MLX team can give them a usable platform.


> surprised ai companies are not making this workflow possible instead of leaving it upto users to figure out how to get voice text into prompt.

Claude on macOS and iOS have native voice to text transcription. Haven't tried it but since you can access Claude Code from the apps now, I wonder if you use the Claude app's transcription for input into Claude Code.


> Claude on macOS and iOS have native voice to text transcription

Yeah, Claude/ChatGPT/Gemini all offer this, although Gemini's is basically unusable because it will immediately send the message if you stop talking for a few seconds

I imagine you totally could use the app transcript and paste it in, but keeping the friction to an absolute minimum (e.g., just needing to press one hotkey) feels nice


Due to not shipping GPL3 code, the version of Bash on macOS is 3.2 and sometimes that makes a difference.

> They just need to start deprecating and removing old features.

That's not going to happen.

They can't break millions of websites by removing old features. Besides, for the most part, current developers can ignore most of the old stuff.

But a site made in 1997 has to render on current browsers.


>On the contrary, a lot of the reason CSS is confusing is because it's full of insane hacks people have to do to get the behaviour they want.

CSS is confusing because the vast majority of web developers never learned it properly. Many developers won't learn any "new" CSS (like CSS Grid which shipped in all browsers in 2017) beyond the hacks they learned in the '90s and early 2000's.

That's not the fault of CSS.


> CSS is confusing because the vast majority of web developers never learned it properly. Many developers won't learn any "new" CSS (like CSS Grid which shipped in all browsers in 2017) beyond the hacks they learned in the '90s and early 2000's.

Disagree. The newer stuff is, if anything, more confusing. The old stuff, awful as it was, at least had a consistent model.


> Disagree. The newer stuff is, if anything, more confusing. The old stuff, awful as it was, at least had a consistent model.

With the "old stuff", we didn't a layout model or an alignment model. Everything required float and positioning hacks to do things they weren't designed to do. There's no logical way that was "better."

There were several different grid systems, all mostly incompatible with each other, which were required to do anything interesting.

Many layouts that are common today were impossible to do with just HTML & CSS back in '90s and 2000's.

Capabilities that almost all developers had to reach for a framework Bootstrap or Foundation for are built-in to CSS today. Or lots of JavaScript.


A gentle reminder: conditionals aren't new to CSS; @supports and @media are conditionals; so are style queries.

if() just codifies behaviors and hacks [1] developers were already doing.

[1]: https://lea.verou.me/blog/2020/10/the-var-space-hack-to-togg...


> Prompt ingested (time to first token) in "18 seconds" with the new chip... end of the joke

18 seconds on the M5; 4.4x times faster than the previous M4 running one of Qwen’s 8 billion parameter local models.

That’s quite impressive for a tablet and faster than most laptops available today.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: