Tuesday, November 10, 2020

Apple Silicon: The Roads Not Taken

Jesper:

Raspberry Pi-like, “tinkerer-friendly” Mac, for under $100.

[…]

A Mac mini the size of the Apple TV, for $199, with 4GB RAM, 64/128 GB of iPhone-like storage, hardly any I/O, and probably an A12, A13 or A14.

[…]

Take the current Mac mini, make it a bit smaller and make it affordable. Again – the Intel tax is gone, and Apple, if they want to, can churn out silicon in large scales by themselves already. The first Mac mini was $499 – there’s no reason the first ARM Mac mini can’t be.

[…]

All of these products essentially are based on this: there’s an Apple that makes iPhones for $399 with industry-leading performance, and there’s an Apple that sells wheels for almost twice that price. It’s up to Apple to define what they want to sell and how they want to market it, and heading into a transition where you drop a hardware partner for your own designs is a perfect time to choose a new tack.

Sebastiaan de With:

Cool performance bump for the MacBook Pro, but today’s PC laptops have fantastic screens with tiny bezels. Face unlock. Touchscreens and pencils. 4G modems. Tons of awesome features. This looks like a faster MacBook Pro from 2015 with a tiny touchbar nobody wants.

Frank Reiff:

Perhaps the most puzzling thing is that they haven’t made a successor to the 12” MacBook. That would have been amazing. The only things wrong with it were the performance, the single port and the keyboard: all things that they could have solved with the M1.

Frank Reiff:

For those who hoped that Apple was going to use the Apple Silicon transition to course correct on issues such as ports, Touch Bar and perhaps introduce a new design language or new iOS features such as FaceID, tonight was a bitter disappointment.

Previously:

Update (2020-11-20): Samuel Axon:

And FYI: we also asked if Apple plans to introduce cheaper Macs, on the assumption that using its own silicon is more economical. “We don’t do cheap—you know that,” Joswiak admitted. “Cheap is for other people, because we try to build a better product.”

The question wasn’t about Apple making Macs that were cheap, just cheaper than the current ones.

Update (2020-11-23): Tanner Bennett:

People need to stop using the word “cheap(er)” when talking to Apple execs. Ask them about affordable Macs.

23 Comments RSS · Twitter

(Wants to pat all the pundits who do not remember the PPC-Intel transition on the head)

Apple bloggers have been saying for months that Apple would probably echo the 2005 transition to intel: No changes whatsoever to form factor, basically switch the chips inside and keep everything almost everything else as same-same as possible.

A Pi-like "Mac Micro" would be an awesome idea.

16 GB for the 13'' MBP and the Mac Mini is joke. But it seems that's as far as the M1 will go.

I think it was clear they weren't going to try any radical new designs at the same time as a brain transplant. There's just too many ways it could go wrong. And when it does go horribly wrong ("The universe tends towards maximum irony"), consumers would inevitably blame the CPU.

No, they were always going to pull a "cheesegrater": keep the same case, and make the CPU part smaller, and the rest bigger. In that case, it was storage and expansion. For the laptops, it's batteries.

Jesper sums up a lot of my feelings about this. There's an opportunity for Apple to make the Mac affordable for many more people, but so far they've chosen margins over 'a computer for the rest of us'.

A MacPi might have helped make Apple relevant in education again, but when their entry-level computer is still $899 for schools (and $999 for everyone else), it's hard for developers to give much time and attention to Apple in that field.

"The first Mac mini was $499 – there’s no reason the first ARM Mac mini can’t be."

Apples and oranges. Taking 15 years of inflation into account, the first Mac mini was $665 in today's money. It's only gone up 5% in cost.

@Ted. And that introductory price for the first Mini was for a computer without wifi or bluetooth. To get something vaguely like the $100 more expensive 2006 mini, you had to buy the upgraded version.

@remmah Pining for a more affordable Mac "for the rest of us" is a nice idea that will never happen. There never was such a computer in the entire history of Apple. Apple has always pursued high margins and sold computers that "the rest of us" could not easily afford, from the Apple ][ era to the first Macs to today. Nowadays, Apple comes as close to that ideal as it ever has - they make computers that cost about $700 to build and sell them for $1000.

To significantly lower their prices, they'd have to either produce pieces of junk that don't work well and break easily (ie, every PC in Best Buy). Or else they'd have to stop being a for-profit publicly traded company, change their entire corporate ethos, and sell their computers at cost.

@Ted @Glaurung I don’t really understand the inflation argument when elsewhere in the industry prices are dropping. What is the evidence that a cheaper Mac would be junk or sold at cost?

Presumably Apple sells the A10X-based Apple TV 4K for a profit at $199. Swap the Siri Remote for more (not eMMC) storage, and I'd imagine they could sell a Mac nano for $299, and laptop with the same parts for $349-399, and still make a good margin.

That said, based on discussions elsewhere, the older A-series chips apparently have some weird edge cases that would make them unsuitable for using in to run macOS as-is. Perhaps over time, the M1 models will be kept around and sold for lower prices.

In the meantime, I'll be working on apps that run on the Raspberry Pi 400.

The argument that lower cost PCs are junk is totally baseless. I just bought a Raspberry Pi 4 which is amazing for $70 and an Intel NUC for $200 (256GB SSD/8GB RAM). Both are very high quality and include an array of I/O (not just one port like the new MBP, WTF?).

The computer industry as a whole has gotten nearly an order of magnitude cheaper in the past decade. Compare what it would have cost to fab 100cm2 of high quality PCB 10 years ago vs now. Or the cost of other parts like SMD chips. You can buy 100 standard SMD resistors from a name brand for 1 cent at retail now, and the factories that build Apple's parts get them for more than 10x cheaper than that. So it's literally true that this stuff costs pennies.

Apple could definitely build a totally capable Mac Mini for $400 (and still have a 30% margin to cover R&D) that serves the needs of a vast array of its customers (and potential customers) and have a long useful life. We are way past the days where computers become obsolete in a year or three -- everything these days is insanely fast. Hell I still regularly use my 2014 MBP for normal tasks when I don't want to sit at my desk and use my PC. It doesn't feel slow at all. Not everyone out there is editing 4 streams of 8K video all day long. Running Safari, Mail, MS Office, and some iOS apps to do Weather and Netflix (now with Big Sur) smoothly would probably cover the needs of a HUGE majority of people.

Hoping Apple would make cheaper computers now that it no longer has to pay the Intel tax would only make sense if one thinks that they are a bunch of engineers and artists who care about product quality and want their creations universally available. (That's what the folks at Raspberry Pi want -- their computers are priced accordingly, and their R&D budget is accordingly tiny).

In fact Apple cares about one thing only: its profit.

Prices are the highest the market will bear. (It is very unlikely Apple pays $200 more for that extra 8Gb, but customers can be expected to pay it since there is no way to upgrade the computer once bought).

Savings from this transition will accrue to Apple, while costs will accrue to 3rd party developers. Since I expect derivatives of the M* series to be reused in iOS devices, Apple's costs actually will go down, since they won't need to support x86 anymore. In contrast, Mac software that pushed the envelope by using assembly, or relied on using eGPUs will need either to transition to another platform, or be rewritten. Products that relied exclusively on Apple's APIs will simply recompile.

Equally, documentation and tools are as bad as the market will bear (long compilation times), while still allowing Apple to control developers by locking them in (Swift/Metal), and locking out competition by providing Xcode for free and changing APIs all the time.

Rent-seeking is the new growth area ("App Stores", expensive cloud accounts, and AppleCare). Their need to control this revenue stream is obvious given that they even want control over apps streamed from third party servers.

Although I despise React & Javascript, moving to web-technologies and making subpar portable software like Slack is probably the least risky way forwards for 3rd party developers. It's not irrelevant that Apple tried to kill the Unreal Engine which is a heavy duty engine used to make AAA titles by large studios. It argued that Unity sufficient -- an engine used to make casual games by tiny outfits. My understanding is that their capabilities differ significantly, and that clarifies Apple's understanding of 3rd party software: secondary experiences, interchangeable commodities, not much different from sitcoms for a cable-TV channel -- the time must be filled, but it doesn't really matter by what as long as enough people watch it. For all I know Apple might be right, and most Mac users might be content with iOS apps... but that attitude doesn't make my heart sing. I'll also point out that when they gave a voice to 3rd party developers, it was only for them to look like experts singing Apple's praises, but not to show off their own creations.

Old Unix Geek wrote: "I'll also point out that when they gave a voice to 3rd party developers, it was only for them to look like experts singing Apple's praises, but not to show off their own creations."

The developer clips were so creepily over-the-top they almost felt coerced. Has anyone checked to see if the devs were blinking out "SOS" messages with their eyelids?

Over the years, many people have joked that Apple can be a bit cult-like, but the live stream really took that to another level. The physical posture of the speakers...so many creepily rehearsed smiles...the slick, yet remarkably stiff, production quality.

Strangely enough, I thought Apple did a terrific job with the WWDC streams, even though those certainly were as stage managed and rehearsed as the "One More Thing" event.

@Michael Tsai - Exactly.

It still baffles me why do people who have no understanding of BOM cost, comes to say you cant design a Mac for cheaper price or risk making it junk. They are reading too much form likes of DaringFireball who claim the $100 AirPod are sold at cost. ( No they aren't , and no offence to John Gruber)

Apple could have sold the current Mac mini at $499 while maintaining the *same* margin as their previous Intel Mac mini. I am speaking of margin here, in percentage, which very often people think of it as Net profits in absolute terms.

Without the M1 or A14X, You could make a 100% ( not 99%, but 100% ) replica of Mac mini at those volume for $100. The 8GB LPDDR4X Cost $32 without volume discount. If you put down the Die Cost on M1, Apple would still have their "used to be" "normal" Gross Margin of 60% selling at $499.

Instead, they charge it at $699. $999 and $1299.

And Oh, no NRE and R&D discussions here before someone jump in. Those are not included in BOM, and should be deducted from Gross as total R&D. ( Apple dont sell those as individual units )

My only hope, is that in the future, after the transition, ( so in the mean time their Intel based product are not impacted ) they will start offering more value. ( They will likely made it a big PR and Marketing play once they do it. ) Such as setting 16GB Memory as Standard. And bumping the Base Storage to 512GB. Mini-LED on MacBook Pro. And offer specific models, with 8GB and 128GB or 256GB SSD for Educational only at a much steeper discount.

And I forgot to add, the old Mac mini runs on market leading 3x3 8021.ac WiFi, the new Mac mini, if I would bet, is the same WiFi 6 from iPhone running at 2x2.

I'm always impressed by anonymous people on the internet who claim to be able to estimate Apple's costs and margins to within 1% or $1, for product categories which have only been successfully developed a couple times in our industry.

Have you ever managed tech support for 25 million developers? How much does it cost to write and verify the microcode for a new CPU architecture? How much does it cost to write and maintain macOS -- or in this $499 fantasy world, are you assuming Apple would go back to charging $129 for each release of the OS? Could anyone put a price on Safari at all, considering even Microsoft gave up trying to maintain their own browser?

The few successful examples we have are so ridiculously different that I can't even begin to guess. Do we count LOC and multiple by industry average cost per LOC? Do we take the dollar cost to develop Windows (lots!) and Linux (little!) and pick the average? Do we add up the amount of VC funding pumped into Netscape and Intel and Microsoft?

The cost of a Mac isn't the BOM, because the value of a Mac isn't in the mass of metal and glass and plastic it contains. You might as well claim everyone should be able to download Hollywood movies for the cost of the bandwidth because they're just bits.

@Ted If you think that Mac hardware margins are what pays for tech support for mostly iOS developers, I don’t know what to say. Development of CPUs, macOS, and Safari are all fixed costs for Apple. So, in theory, they could be amortized over more Macs if Apple lowered its hardware prices and margins. That may or may not be a good idea, business-wise, but it’s not one that can be just dismissed, especially given Apple’s recent focus on services revenue. I think some back-of-the-envelope calculations would show that the long-term cost of Apple Silicon in Macs will be lower to Apple than using Intel chips and that the software stuff doesn’t add up to a lot. Plus, in both cases, much of the work is already being done for iOS.

I'm always impressed by anonymous people on the internet who believe two trillion dollar corporations spontaneously arise without a ruthless focus on profit margins.

The way you compute the cost of a CPU or a piece of software is simple: how much did you pay to create it? The market value is whatever people are willing to pay for it.

Do those numbers actually reflect the true worth and cost? No: Apple didn't invent most of the technology it uses. They, like everyone else, or actually more than anyone else stand on the shoulders of giants. It would be nice if they remembered that more often. Safari's origins lie in WebKit/KDE. Darwin in BSD. "Apple silicon" in ARM, Imagination Technologies, Xnor.ai, etc. But the attitude they profess is "We're the best! F*** the rest!" That happens to rankle me, particularly when normies take it as gospel truth.

Isn't Apple *exactly* the type of company that COULD easily make lower cost products, simply because they already do sell so many super high margin ones? Otherwise, what are those $700 wheels, $1,000 display stands, and $100-RAM-upgrades-sold-for-$600 paying for?

Person who worked at Apple (albeit in one of their stores) says:

It’s dangerous to fall in love with a company. When Steve passed in 2011, Apple’s culture and leadership began to change. Instead of a relentless focus on product innovation and customer delight, priorities shifted to maximizing growth and extracting more revenue per user.

https://hardware.substack.com/p/falling-out-of-love-with-apple-part1

I'm always impressed by anonymous people on the internet who talk about Apple's margin and business without knowing anything about Accounting. No one ever said Apple would sell it at cost. But it is quite obvious people dont have a clue who R&D and NRE are deducted. And have no idea how economy of scale works.

And yes, the only unknown variable in the BOM costing is M1. Everything else, the parts forgive me, are commodity.

And in case people dont know, as part of their accounting change in 2018 Apple currently put $10 per unit sold towards services revenue as Software development cost including but not limited to Siri, OS and Maps.

Long gone are the days of “Do you want to sell sugar water for the rest of your life, or do you want to come with me and change the world?” Now, it's all just "commodities". Who cares about tech? What really matters is Accounting. Send your best and your brightest there.

What a depressing world you paint, Ed.

Old Unix Geek November 12, 2020 7:39 PM shared "When Steve passed in 2011, Apple’s culture and leadership began to change. Instead of a relentless focus on product innovation and customer delight, priorities shifted to maximizing growth and extracting more revenue per user."

I just spent some time trying to remember the last time I was truly delighted by a new Apple product. While I was satisfied by the mid-2012 MacBook Pros, I was delighted by the iPhone 4S.

The iPhone 4S shipped nine days after Steve Jobs died.

Apple spends 2.7% of its revenues on "R&D". (That includes software development.) It's less than most of the rest of the industry.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.720.3278&rep=rep1&type=pdf

This study is well worth reading. It makes it clear that many of the things I intuited and have been bothering me were correct. For instance, Apple purposefully creates barriers to customer exit.

Old Unix Geek: Apple spends 2.7% of its revenues on "R&D".

2.7% of what?
I'm sue it surely be a lot more than GDP of lots countries.

Leave a Comment