321–330 of 338 posts
Earlier quoted context omitted.
Maybe. But what are the options? Windows? Not really. Linux/Debian/Ubuntu is great, in theory, but it lacks tools: Acorn, Keyboard Maestro, OnniGraffle, Alfred App, MS Office, MS Teams, and a dozen more. Plus Apple silicon. I test LLMs locally.
Platforms that lack MS Teams are doing the user a favor.
Earlier quoted context omitted.
Nope, on cloud it is faster than conversational where you can skim to the right place or quickly regenerate seeing it is going down the wrong path.
I take that to mean that it is basically so slow as to be useless and that you would in fact need a Macbook to be able to run this model locally. This whole thread has become a bit frustrating with comments about how other laptops are competitive with a Macbook. But no one can actually name a single laptop that can do all the things that a Macbook can.
For desktop I'd still rather threadripper etc. where it has more competitive CPU memory bandwidth and is upgradable and can run multiple GPUs.
Earlier quoted context omitted.
Maybe. But what are the options? Windows? Not really. Linux/Debian/Ubuntu is great, in theory, but it lacks tools: Acorn, Keyboard Maestro, OnniGraffle, Alfred App, MS Office, MS Teams, and a dozen more. Plus Apple silicon. I test LLMs locally.
And here I am running Linux on M1 Macbook. Graphical interface is superior, native docker support is superior, development experience is superior, backups for sure are superior than Time Machine. For work purposes online office on the web is sufficient. It is not for everyone (like photographers), but for software developers it works very nicely.
i.e. a device directly connected via ethernet can be communicated with when running the binary normally, but cannot under lldb. Meanwhile no prompts about allowing network connections etc. are presented.
That wasted at least half a day until I realized I could just use the upstream LLVM debugger.
Earlier quoted context omitted.
Apple has 10% of worldwide desktop market, and no presence on servers and cloud. They aren't eating anyone's lunch, not even on mobile, where Android rules world wide. Microsoft also owns the PC, besides XBox, Valve is a guest and should not forget how many studios are owned by Microsoft. And with a Microsoft friendly US administration, maybe even Valve-after-Gabe can become part of Microsoft.
Apple owns 10% of the market, but around 90% of all computers above 1000$. They make around 7.5 billions in Mac revenue per quarter, which makes them the most profitable computer company. Apple controls the high margin market. That has royally screwed Intel, which explains in part the woes its in right now; it lost the only high margin computer maker.
Earlier quoted context omitted.
The day Microsoft decides Proton is a nuisance that gaming alternative is gone, as Valve decided fostering a GNU/Linux game ecosystem wasn't worth it.
I'm not quite sure what you are saying here but Valve can beat Microsoft at almost any game.
And apparently the only way to get games on GNU/Linux is to build on Microsoft's work with reverse engineering APIs.
If they can beat Microsoft so easily, where is Valve's own platform, without Microsoft dependencies?
Earlier quoted context omitted.
> to the cloud to locally Requires payment for backup of any meaningful size. > locally to your own storage I don't own a mac. I own Linux. How do I backup iOS to my device? How do I restore it?
> I own Linux Does idevicebackup/idevicerestore not work anymore? But agreed, there's no technical reason why the phone can't just accept normal USB-connected storage and store a file in there (in whatever proprietary/encrypted format it wants, but at least the _storage_ is standards-compliant).
Earlier quoted context omitted.
It’s very easy on iOS to make a backup of your phone including all apps, settings, and data. Either to the cloud to locally to your own storage. What am I missing?
Cloud does not work. If you have an app that is no longer available in the App Store, when you install from a backup you are out of luck because that app can’t be downloaded from the App Store. If you choose to manage your phone from a Mac, then you can save a full backup.
Earlier quoted context omitted.
A point is being missed here; The user, nor the provider of hardware/software can be hostile towards each other. Here's why: The cloud is someone else's computer, not yours. It's a shared drawer that you get a key to, but so do many others. If our main access to your drawer is through the programs on our computer, it also can be limited and broken for hours, days, or weeks while under repair. I'm not sure how many la…
There are many difficult technical and other important considerations and trade-offs and I think that Apple is answering them the majority of them as best it can. There are major benefits to having the kind of integration that Apple has, for example, the security chip and the wifi obviates theft of laptops and phones. You cannot get that with the traditional "PC starting from a boot partition" model. I lost a lot of…
The simple thing is choice. A switch to turn it off. By making it locked down, it also does make it more insecure.
Open Source and auditable code is better than private code that no one has an idea of whether it's compromised or not, and for how long.
The recent iPhone zero days are an example.
Earlier quoted context omitted.
> Apple is about the only company that allows you to run any decently sized LLMs on a laptop Comments like these make me wonder if people even use different computers anymore or if they just say stuff and hope that it's right.
Apple is the only company that makes integrated GPUs that are actually any good, and even their CPUs are really well-optimized for these sorts of workloads. They also build their RAM in a way that makes it suitable for both CPU and GPU uses. This means you don't have separate CPU RAM and GPU VRAM, it's all just memory. If you get more RAM for your Macbook, that automatically means more memory you can use for LLM infe…
I will reiterate; if you have used AMD or Intel iGPUs recently you would know this is just plain wrong. AMD's iGPUs often outperform Apple's for the price, and come with a featureset that isn't intentionally gimped and supports "professional" features like Vulkan conformance. People condemn Apple's GPUs for having a featureset too simple for anything besides video transcoding and compute shaders. If Intel didn't exist, Apple would probably be the single least-efficient and lowest performance GPU designer shipping products today. And they had to cut corners to get there.
> If you get more RAM for your Macbook, that automatically means more memory you can use for LLM inference.
And then it's your paltry GPU that's the bottleneck. You can get comparable results on pretty much any sub-120b parameter model with PCIe offloading on Nvidia hardware that costs a fraction of a Mac Studio with enough memory to compete.
See for yourself. Nvidia's fastest laptop chips still outperform Apple's most powerful desktop GPU chips: https://browser.geekbench.com/opencl-benchmarks
And it's not even close. Apple's biggest professional GPUs struggle to swap spit with Nvidia's last-gen laptop chips.