Comment by jrockway | Hacker News Reader

Comment by jrockway | Hacker News Reader

Apple's chips are really impressive. I would love to see them in a general purpose computer so that developers can really take advantage of the power. Whatever is in my iPad Pro would make a great development workstation if I could connect two monitors to it and ... run Linux. It is power-efficient, very lightweight, and has built-in 4G. In those categories, it beats the pants off of any Intel-based laptop I've ever had. But I still have to SSH to a real computer to work, which is an artifact of Apple's OS policies, not what sort of workload the chips they put in there are capable of.

The unfortunate reality with all of these stripped-down computers, though, is that nobody is writing software for them. While the CPUs are powerful enough to run pretty much anything, nobody is porting their AAA games or CAD software or video editors. Maybe Apple's ability to dynamically recompile Mac Store apps to ARM will help, but I still think MacOS is seriously lacking compared to Windows. So the best Apple can hope for is a "budget Mac" which power users won't buy (and power users write all the software), and they'll just be an expensive Chromebook. (ChromeOS doesn't have a lot of native software either, but at least you can run a normal Linux environment in a container and write software on the device. And, I don't think any person goes out to buy a Chromebook... they just get them for "free" from their school or company. It is an interesting segment of the market that can be profitable, but I really want to see these chips in something I can sink my teeth into. Set up a development environment, write software, and run the same software in production on those chips too.)

My takeaway is that it's a shame that this amazing CPU-design division exists inside a company that is all about dumbed-down user experiences. They could probably kill Intel overnight if you could just buy the chips on their own, but instead they insist on putting these chips in devices that are hamstrung by arbitrary policy. Obviously they are playing a long game, but as someone who just wants the best possible equipment... I can't afford to wait. (And I don't think the end result of the game -- Apple makes the CPU, laptop, OS, and software -- is very good for anyone.)

The whole point of these rumors is that Apple's putting these chips in MacBooks, running macOS. So I don't get this kvetching about "stripped-down computers" and "dumbed-down user experiences".

We've had a divide for over a decade now, where serious computing is done mostly on x86, and mobile mostly on ARM.

ARM devices have become associated with App Stores, touchscreens, locked-down and dumbed-down systems. And there's a very real risk that Apple will take this as the opportunity to completely lock out any software that's not been censored and taxed by their App Store.

If they are going to lock down macOS totally they might as well kill it and exit the pro market entirely. If I can't run whatever I want on it, it is not a real computer.

I personally doubt they will do that. If anything I could see them making iPadOS more capable, moving more of the consumer market there, and making the Mac line even more suited to pro users with more compute power and more OS capability.

BTW the ARM move could fit into this. I wonder how a 32 or 64 core A series ARM64 chip in a laptop form factor would perform? Our company runs some infrastructure stuff on Cavium ThunderX chips with 96 ARM cores and the performance is comparable to high end Xeons for very parallelizable load. The Apple A series core is far superior. I'd take 64 of those in a laptop.

It would be so fast I'm sure it would have no issue running transpiled X64 code at perfectly adequate speeds. Could probably even run X64 VMs in emulation at speeds that were at least quite usable for testing and dev.

Do you really suspect that Linux would run as well as iOS? Isn't iOS optimized for the A17 and vice versa? It's out of my wheelhouse, but I suspect that this synergy is the only reason they seem so fast.

I am going by compute benchmarks, which should be relatively unaffected by the OS. (The OS could slow them down if it really wanted to, I guess.)

Random example: https://www.macrumors.com/2018/11/01/2018-ipad-pro-benchmark...

Those benchmarks don't seem to account for batteries.

I hook my 2018 MacBook Pro up to a 4K TV to present to a meeting and the battery goes down to about 35% after two hours.

2 hours is a reasonable amount of time for something like this.

We could optimize things for iOS if it were easy to run. Has happened for dozens of platforms.

From recent experience using the Pinebook Pro and manjaro, the answer is a resounding yes.

From the start, it was clear that a decent arm chip in a laptop is very usable. Paired with an SSD, and more ram, it would be killer.

MacOS would need some care to make it less heavy in terms of background consumption, but I don’t see that as being beyond Apple’s reach. Windows is a different matter, not sure what MS can bring to the table there.. The work being done right now in porting iOS applications to work on macOS clearly points to a strategy being developed.

Think it’s going to be an interesting couple of years.

Nobody is porting their big pro software to the ipad because it is ram-starved. Some things just need ram, lots and lots of ram.

Why don't we see more OS-level integrated remote computing in the Chromebook category?

I get Google started it to run their webapps.

But it doesn't seem technically infeasible to package and sell a set monthly subscription price, that gets you a lazy-allocated machine on GCP.

Spin up beefy software, your computer recognizes "this is more than I can do", checks current networking availability, and starts running the app remotely.

I know there are offerings like this, but it seems like a killer feature if it had full OS integration.

Just plop an overlay icon somewhere in the window to clearly note it's running remote. (Thereby letting users reason through why it occasionally behaves weirdly)

People don't have room for any more subscription fees in their lives. Especially to pay for something like this (or game streaming) that offers an inferior/laggy user experience, when owning a reasonably powerful laptop or games console isn't all that expensive.

Subscription fee exhaustion seems an odd concept to me. Do people actually say "A service delivers me value for its price, but I already have too many subscription services, so I won't use it"?

It seems an ironic perspective given that "Time is the only real currency we have" is on the HN front page!

I specifically mentioned it because Apple (1) has cloud resources, whether they see it as an offering or not, (2) has full stack integration with their products, & (3) has historically prized ease of use and packaging new technology in a slick way.

"Underpowered Mac / iPad / etc, now with iPower transparent cloud compute" seems a unique way to leverage their strengths and abilities in ways that Google, MS, and Amazon lack the capabilities to match.

> Do people actually say "A service delivers me value for its price, but I already have too many subscription services, so I won't use it"?

Sounds like Silicon Valley thinking. If you're used to massive salaries (and massive housing costs), you probably won't care about relatively small subscription costs.

But for the majority of people elsewhere in the world, those costs really do add up. There's very little appeal in renting a PC or games console in the cloud, then getting a sub-standard PC or console experience for an ongoing fee.

Just to add to this - life is pretty expensive these days, and a lot of people really don't have much discretionary income left each month after taxes, rent, health insurance, food, etc. Median household income in the US is ~$63k, and after all of those expenses its pretty easy to be left with something like $100 a month (or less, or even worse, needing to take on debt to cover bills). When that is the case, even a $5/month subscription is a lot of money.

Point not being that subscription is not a lot of money.

Point being that expecting someone with that level of disposable income to having savings that allow for capital expenditures (instead of amortized subscriptions) isn't likely either.

It's about input devices and UI features much more than RAM.

iPad apps are expected to have massive buttons for big fat fingers. Pro software requires a mouse and a keyboard, a much more precise and flexible combination.

(Yes, you can get the Apple Pencil and an iPad keyboard. But it's a bit chicken-and-egg, people don't have the accessories because there's little software that uses them. And there's no software as few people have the accessories...)

Everyone I know with an iPad Pro has the keyboard case. I hate the keyboard... but I can put up with it while travelling. I don't think the keyboard availability is an underlying issue, but I do think that people will balk at a $700/year license for something that runs on an iPad, because they feel like it's a toy, not a tool. However, the CPU is not what causes that; the CPU has more raw compute power than many Intel laptops that people do run expensive software on.

I want my iPad Pro to be a good workstation, but the reality is that will probably never happen. However, if Apple uses the same hardware to run MacOS, then it could happen. Yes, they would need to tack on some more I/O (or at least release a first-party dock) and RAM, but they could still make a very thin and fast device that would blow Intel laptops out of the water. I would not buy one of these because I don't really care for MacOS. (I use the iPad as my "laptop" because it doesn't really require any customization or maintenance. I have my SSH key on there, that's about it. I use it as a dumb terminal and it's great for that, but with the right software, it would also be a very capable computer.)

Have you looked at the iSH beta? https://ish.app/

Reply on news.ycombinator.com