Apple silicon is winning in performance and power consumption. It's not quite the same as the old Intel based generation, where they were more like premium/well built machines with good screens and hardware.

Now some people would like to use that top of the line Apple silicon but without MacOS that comes with it. I doubt this is a huge threat to apple but it's also not a huge market for them. From what I understand they are not actively impeding it but not actively helping it either.

I actually won a M2 laptop a while ago. I resold it, new in box still shrink wrapped because I did not care for MacOS, just a personal preference. People like me are a minority though.

I'm in the same boat.

Hell, I was given a bunch of old iphones recently, all of which are newer and better hardware than my phone. I wiped them and donated them to charity.

Intel have a chance to be relevant today, but they couldn’t imagine a future with smaller better faster, computers the same applies to IBM and Motorola of Schaumburg, Illinois. They just couldn’t imagine anything beyond.

Intel was asked to design a chip for a new smartphone. They said no, which was the end…

Note: Intel and Microsoft have trouble competing when the playing field is level.

What is the purpose here? You are still stuck with a full blown MacOS running the system no? That is what Asahi tries to replace, not the other way around.

I'd recommend reading up on it, and perhaps trying it, because it's very useful.

For folks lamenting that the (awesome) Asahi efforts are challenged by reverse engineering of changes from M2 to M3, and so on, those changes vanish into the abstraction, but their benefits do not. That's why i switched from Asahi a couple years ago to UTM in apple silicon hypervisor mode.

https://developer.apple.com/documentation/hypervisor https://apps.apple.com/us/app/utm-virtual-machines

The point is some people who want to run Linux mostly do not want MacOS. You suggest to "just run MacOS" (the thing they want to avoid) and virtualize Linux. That defeats the entire purpose, then you could just as well skip the VM and run Mac apps. Yes, there are a few specialized applications people might need Linux for but that's mostly not why they install an entirely different OS. For the occasional obscure programm that doesn't exist in the Apple ecosystem VMs are a good option, absolutely. Everything you do in the VM still runs on Mac though and you depend on Apple for security, privacy and trust in general.

yeah i get it, and run native linux myself, but also have apple silicon. for some, but not everyone (to your point), the apple hypervisor.framework is itself an adequate trick.

It serves the exact same purpose that WSL does in Windows.

You can run any flavor of Linux you like in a virtual machine, while booted into the native OS.

The purpose is to be able to run a decent Linux setup while also having MacOS for things. Not everyone is stuck on the anti-Apple train; some of us put software on all the platforms and thus using the hypervisor for this use case is quite comfortable. Disclaimer: personal experience - I too use UTM for highly productive things..

Wanting to run a single OS at a time is anti Apple now? Run MacOS or run Linux, why run both and create more problems than it solves and waste resources?

Run both when you need to build for both. Why not run both when you need to?

yes, but you're arguing for running both whenever you need (only) linux?

I’d need both, so I’d use both.

I must have missed something but can't you just prompt it to answer in your desired style? What am I missing here.

Commenting because I am struggling with this too, claude code seems to be so verbose no matter how I prompt it.

You can prompt it, and it almost immediately forgets. Or context compaxtion purges ut. Or it somewhat communicates with you in it, but writes all comments and docs in its original bloated style.

Training supersedes random markdown files and prompts

No, training tends to override system prompts at lot of the time.

Let's say Claude Code's system prompt is updated to recommend responding in Simplified Technical English. It'll work for the current generation of models. But Anthropic will train the next generation via RL on Claude Code traces as they do today. As long as they don't change their reward design, the next generation is going to be pulled towards Claudish again, because speaking Claudish gives higher rewards, so in the end the prompt doesn't really matter.

Ideally these things shouldn't work like this but this seems to be the sorry state of RL right now.

You miss that it is not just your prompt but also the various system prompts, plus how the model was trained. But you can reduce the verbosity (also with a setting in /config).

And memory and code comment styles - i think that’s a big one people forget about. You can prompt it all you want, when it sees elaborate comments in memory and code it will follow the style

But you can prompt it to reduce the verbocity of the comments as a project in itself - I will probably try Fable 5.1 for this.

Training is far more important than user prompts

This. I was looking for a quick Dart board score tracker app to keep scores for family games.

The play store results were ATROCIOUS and user hostile, forcing account creation, full screen video ads and other BS.

I remembered I had f-droid and I found a perfect, no bullshit app that does exactly what I needed and no ads.

When testing Claude code in auto mode in a fresh sandbox with a docusaurus website freshly cloned, I asked:

Can you see the docs folder with the git project?

It was in auto mode. So it immediately saw a docusaurus site (good) but instead of stopping there, it installed nodejs from a static binary download (no root access so only way), ran npm install, started the dev server and confirmed the project worked.

That is some crazy amount of leeway for an intent based classifier. I'm not surprised it's full of holes, and it seems to be 100% by-design.

Fair. I can be overly precise in my communication style, let me know if you'd prefer a more delicate or load bearing tone. Just say the words.

Debian is notoriously very strict on accepting maintainers and they are not anonymous.

tl;dr: stripe sees tokens as the next currency and chose to buy their way in.

It's possible using overlay networks but other than that there has been no progress on that front. The routing table updates would overwhelm the system. Ipv6 does nothing to address device roaming. There is multipath TCP but I have never encountered anything using it.

Most improvements on that front has been happening on the application layer, where the app attempts to retain functionality even though the underlying network conditions are changing, to various degrees of success (mostly not that great).

There's a great hyper-sci-fi RPG setting called Numenera, set a billion years in the future, that has this cool little concept of a "datasphere", as in analogous to atmosphere, an omnipresent permanent connectivity network..

Shouldn't that be our end goal with all this tech?

> There is multipath TCP but I have never encountered anything using it.

IIRC siri uses it and iOS (maybe macOS too?) exposes it to app developers.

Wait till you see the number of actual chip vendors.

There is about 3.. Broadcom, Qualcomm and MediaTek. Together they control I believe 100% of wifi chips used in routers and APs. Often it's sold as a package / system on chip so it's also the main cpu.

On the client side there is a little bit more competition but not much.

Client side don't you just add Realtek and Intel? I don't think almost anyone else does wifi chips.

Apple makes their own connectivity chips. It was a big blow for Qualcomm.

Yup. I might have forgotten Marvell in the list but yes, that is pretty much it.

I am also not sure of their exact market shares but I would suspect one or two are probably steering the majority of the "standards". I'd suspect Broadcom and Qualcomm to have the most influence.

I believe Intel is also a pretty prominent member in the standards committee. Mostly for historic reasons as I don't think they have too much market share outside of PC/laptop wifi modules.

But I agree, Broadcom and Qualcomm are the heavy hitters.

Realtek wifi chips suck, their drivers are always buggy