Claude at Apple
Apple recently announced an AI partnership with Google. But reporting indicates the company initially pursued deals with other companies, including Anthropic.
Based on new comments from Bloomberg’s Mark Gurman, it’s easy to see why.
Gurman, speaking on TBPN, said the following:
Apple runs on Anthropic at this point. Anthropic is powering a lot of the stuff Apple is doing internally in terms of product development, a lot of their internal tools…They have custom versions of Claude running on their own servers internally.
Aaron (Hacker News):
Apple accidentally left Claude.md files in today’s Apple Support app update (v5.13)
That’s an odd way of phrasing it, because it makes it sound like the file is naturally in the app package and Apple forgot to delete it. But why was it being copied into the app in the first place? (It seems to be related to building the app, not to using AI for customer support.)
Apple has released an emergency update to the Apple Support app (v5.13.1) to remove the Claude.md files
I have a friend in apple.
He has over 200 dollars credit on claude everyday to spend.
Previously:
15 Comments RSS · Twitter · Mastodon
Does all this use of LLM tools (both by Apple corporate internally, and by users of Apple devices) count against Apple’s 2030 net-zero climate pledge?
Thoughts and prayers to the next generation of developers, who will have to clean all this shit up and have no mentors left to guide them.
@Gord >Does all this use of LLM tools (both by Apple corporate internally, and by users of Apple devices) count against Apple’s 2030 net-zero climate pledge?
Yeah, they're just burning the entire ecosystem for fun at this point, it seems. Capitalism has entered its "runaway train. No one with the power will even try to stop it from crashing into the orphanage" stage
If AI has become so good, why don't they use it to fix all the millions of bugs reported via the Feedback Assistant?
Apparently, humans at Apple do not really care about fixing bugs. So this could be an interesting use of AI.
They could also use it to create the release notes for OS updates.
$200 credit a day? Apple should be embarrassed by this story. They are supposed to have the best developers in the world working there they need this crutch? I wonder how long have they’ve been doing this.
They can’t even make their own model trained specifically for their own frameworks that’s good enough, apparently.
I'm curious, though. Do the people who took a principled stance against Acorn ("they became dead to me") apply that same standard to Apple? What OS are you going to use?
I've been halfheartedly looking at some sort of Linux, coupled with a Framework laptop, but ugh. I don't wanna go down that road even though I know some of the distros have improved a lot lately.
> What OS are you going to use?
I put Ubuntu on some of my old Macs but I don't really like it and couldn't see myself using it as my primary OS. If https://ravynos.com is good when it is released I might give it a go (not because Apple is using Claude but for various other reasons).
"I'm curious, though. Do the people who took a principled stance against Acorn ("they became dead to me") apply that same standard to Apple? What OS are you going to use?"
Second rule in Church of Apple: Anything Apple does is super duper.
(First role is that Jobs is god and no other gods etc)
We use both Claude Code and OpenAI Codex at work, and they’ve helped us ship higher-quality software way faster with a backlog that stays burned down.
LLMs potentially can cause unrelated problems further down the line, but if the Anthropocene is ending in my lifetime because of them, I’d like to have it end with high-quality software.
I’m merely disappointed that Apple doesn’t appear to have a similar deal with OpenAI; the generally-best model tends to switch every few months as one company leapfrogs the other, and even the generally-not-quite-as-good model will catch bugs the “best” one misses.
Quoth someone:
If AI has become so good, why don't they use it to fix all the millions of bugs reported via the Feedback Assistant?
Ever heard that adding manpower to a late software project makes it later because onboarding the new guys takes time, and the new guys won’t become productive immediately?
Same thing goes for AI use. You have to figure out what it’s good at, what it’s bad at, and how to structure things so an LLM can determine itself, on its own schedule, when a change is an improvement. This involves setting up a lot of testing stuff, and I hear Apple doesn’t do much in the way of automated testing (unsurprising for a company that has lots of GUI apps), so they’re starting flat-footed.
> What do you dislike about Ubuntu?
Dislike might be too strong. I just don't really like it. It's good enough I think to keep a Mac alive after Apple drops support for it, hand it to my kids but I always miss the Mac dock being able to see minimized windows, menu bar behavior, and the macOS stuff. Sure I bitch about macOS getting worse but when I'm on other OS's I always miss these things.
I miss Quick Look. I know there is some Gnome sushi thing I always install to get it, maybe I can tweak Linux enough to like it but I don't really enjoy spending time tweaking Terminal commands to get things just right then some OS update might crush my tweaks and then I gotta do it again.
One nice thing about Linux is that updates usually don't screw up custom configs. Stuff like the window manager are meant to be changed by the user.
