Apple's chip lab, home to the most 'profound change' in decades | Hacker News Reader

Apple's chip lab, home to the most 'profound change' in decades

cnbc.com

54 pointsretskrad78 comments

I have so much respect for how Apple does things. The tight integration of hardware and software is not only better for the products and consumer experience, it also means higher margins for the company. There are not many companies that dare to take it so far. I was disappointed to hear that the baseband modem project has been killed off, but I suppose you've gotta pick your battles.

It's not confirmed that the modem has been killed off. Just rumors.

Can any electromagnetic (?) engineers here speak to the difficulties in creating these chips/antennas? Surprising to hear about Apple possibly failing at an engineering project for something that already exists, given their resources and access to talent. If true, could it be difficulty in creating new tech that gets around existing patents? Trying to achieve current results with less power? Unable to compete with the economics of just buying existing tech from competitors?

It's more than "just" a hard technical issue, which are complex but managable RF issues, a gajillion optional modes, complicated corner cases and compatibility with buggy/noncompliant widely deployed in the field.

Apple has to do this without stepping on a welter of Qualcomm patents amassed over decades as the standards developed.

Qualcomm's CDMA patents predated the standards setting process, so was a special case of Qualcomm having the undisputed ownership of a technology that was used in a subset of carriers in the US, for example, Verizon.

However, now that all the carriers have moved to 4G and 5G, which are open standards, Qualcomm can no longer use patents to stop others from competing on a level playing field.

Now they compete on the merits, and companies like Samsung, MediaTek and Huawei are all shipping their own standards compliant modems.

I would expect to see Apple iterate on the tech until they have something they think is good enough to ship.

I think you're referring to the protocol patents which are under FRAND anyway, so Apple can use them (just have to pay). E.g. You can't modulate a certain way without violating a patent, yet the standard requires modulation in that fashion.

I'm considering implementation patents, such as "use a sliding window to filter/damp the amplitude variation due to user wiggling their hand while close to a wall to reduce the amount of re-analysis to change the the best path to the remote antenna"

(BTW I just made that example up for expository purposes)

When the standards setting body is creating a new version of the open standard, your patent cannot be included unless you make a binding commitment up front to license that patent to any interested party without price gouging.

If you won't make that commitment, they use someone elses tech.

Companies simply won't use Qualcomm proprietary tech anymore after having been burned in the past with something they had to maintain for the sake of backwards compatibility.

For instance, there were zero companies on the planet that bought into Qualcomm's recent tech to send text messages via satellite.

> Essentially, the project is dying because Qualcomm couldn't get a single Android manufacturer to add satellite messaging to a phone. Qualcomm's satellite solution didn't require much in the way of new hardware, so the rejection was apparently due to Qualcomm's design of the feature...

Qualcomm says smartphone makers “indicated a preference towards standards-based solutions” for satellite-to-phone connectivity, a plan the company now wants to pivot to.

https://arstechnica.com/gadgets/2023/11/qualcomm-kills-its-c...

This has been explained multiples times before Apple even acquired Intel's Modem team. (It is somewhat similar to how Intel 's GPU competing with Nvidia. )

We are edging close to 2024. So some of those notes are nearly 10 years old and they are still true.

Not only on what Gumby have said. ( Apart from the part Side Stepping Patents ) The price on what you are paying for the best Modem compared to the best CPU is tiny. Intel even complained about Qualcomm's modem being so cheap after discount they couldn't even compete on cost, let alone price.

And of course hypothetically Apple could make a better modem, but if it cost them twice as much over 5 years. One may ask is it really worth it? At least Steve Jobs thought it was a good price to pay to Qualcomm when he stuck that deal. Just like how he pay a lot more for the Corning glasses that most now take for granted. Only Tim Cook wasn't happy with it. Also worth mentioning Qualcomm spend the highest percentage of revenue on R&D.

Basically we are now able to have hundreds of people all doing FaceTime at the same time in the same place over a single 5G cell tower. And most if not all working in tech give zero appreciation on how far the Network has evolved.

We don’t actually know that it’s true, that was a rumor from a somewhat spotty source.

It also seems very unlikely. That would get Apple out of having to deal with Qualcomm and paying them so much.

Apple is willing to spend a lot of money for that kind of independence. Even if it takes a very long time.

Apple has already been caught forwarding Qualcomm's restricted documents to Intel:

https://www.nytimes.com/2018/09/25/technology/qualcomm-apple...

The previous Intel 4g modem performed poorly. Equivalent iPhones with Qualcomm modems (for Verizon) were purposefully reduced in performance to match the bad part obtained from Intel.

"The chipmaker claims that Apple also threatened the chipmaker to prevent it revealing it throttled Qualcomm-powered iPhones. A minor controversy erupted last year after tests found the iPhone 7 from Verizon could download data faster than the iPhone 7 from AT&T, but didn't."

https://www.zdnet.com/article/iphone-7-throttling-qualcomm-s...

Mediatek sells 5g modems. Taking these on as a secondary source seems cheaper, and would cause less legal risk. Apple could buy 20% of the company if they wanted design influence.

https://www.mediatek.com/technology/5g/5g-modem

> It also seems very unlikely.

Agreed. Apple long ago learned that they could plant rumors rather than just be subject to them, and I'd bet Apple planted this one to cool the rumor mill on this topic since they know they're still several years away from shipping homegrown solutions.

> The tight integration of hardware and software is not only better for the products and consumer experience

Not really. This just means you're overpaying for both the hardware (macs are famously overpriced for what they are) and software (apple price gouges both users & developers left and right).

If you're part of the privileged elite with a fat wallet, then sure. It's great.

> it also means higher margins for the company

Why would you respect that? You enjoy being fucked over by bigcorps?

Macs have a lot more value than most Windows PCs.

Build quality, *nix terminal, speed, quietness, coolness, battery life, etc.

And last but not least, resale value is excellent.

I can guarantee you that any mac you will find will be much slower and have far less ram than my desktop PC, and the mac will probably cost at least twice as much.

You have a gym membership? You’re going to need that strength to carry around your desktop, monitor combo.

I have an office.

Me too.

This is what I thought, but then I got a PC for gaming. Now I use it for pretty much everything. It has outperformed and outlasted every single Macs I have owned at half the price. I don't need to resale since it is very easily maintenable and costs a lot less, I can just retire it or give it... I still have a Mac Mini but I have no will to ever change it.

> The tight integration of hardware and software is not only better for the products and consumer experience, it also means higher margins for the company

And yet they feel like they need to price gouge and charge a shit load of money and on top of it do basically the absolute minimum software testing and constantly with almost every update release a TON of new bugs they should have gotten rid of during testing.

You have too much respect for a company that USED to make great products and takes advantage of customers. Apple and Timmy are probably laughing at you.

> "...pretty much all Macs are capable of running Triple-A titles, which is not what it was like five years ago.”

It's interesting that Apple keeps making this point, and that they released the Game Porting Toolkit. I don't think we're going to see like an Apple Switch device, but they do seem to care about console/PC-quality games in a way they never have before.

Games have reached IP prestige parity with movies, music & books. Before they were perceived as some combination of ’geeky waste of time’ and ’unhealthy obsession’.

Still are…

Not sure what you mean. People like me (40+ male) are a huge segment and plausibly spend much more time in games than movies and music.

By revenue games are possibly bigger than movies and music combined (global revenue).

Not everyone needs to like games buy many do and the industry revenue is quite massive compared to any other entertainment.

Still perceived as 'geeky waste of time' and 'unhealthy obsession'. Just an economically successful one.

That sentiment is strongly waning, imo. Chess was similarly dismissed at one point when it became a fad. Gaming is fairly mainstream nowadays. But probably depends who you ask.

The thing is they probably will run triple-A titles but how well and how modern a triple-A title?

Starfield, Cyberpunk 2077, GTA 6 when it comes out? I doubt it.

> I don't think we're going to see like an Apple Switch device, but they do seem to care about console/PC-quality games in a way they never have before.

I think we see this sentence said every few years (or at least a similar one, related to iOS). It'll slowly be as evergreen as "the year of Linux desktop".

Meanwhile, the amount of games that actually are available and running in ARM64 mode on macOS is still miniscule.

Honestly, Game Porting Toolking doesn't seem like a product built by company that's serious about gaming - it's a developer tool that still needs the developer to do extra work to actually publish the game for Macs.

Compare this to approach Valve took on Steam Deck and you'll see just how a company that takes games seriously works.

Exactly. Game Porting Toolkit is an awesome showing, but the intention is in the name; it's not a compatibility layer, it's a porting tool. Eventually it will go away, and you'll be stuck with the same Mojave/Catalina-style cutoff for all your old games.

Apple could make things so much easier by just shipping MacOS with a native Vulkan driver and letting the community maintain the DXVK translator like all the other GPUs. It wouldn't diminish the demand for Metal (it's still the only option on many platforms) and it would bring Mac back up to parity with other OSes. Maybe a renewed commitment to supporting Rosetta would also help seal the deal.

Things like Proton aren't something that shipped overnight, it's a teething process.

In fairness, Valve originally did take the same approach. Proton was born from the ashes of Steam Machines - Linux boxes that developers had to create a port for.

Similarly, WSL, WSL2, WSLg, and WSA are born from the ashes of Windows Phones and Project Astoria. They aren't quite there yet, but I expect another shot at mobiles in the future.

Apple will have to go through the same teething process. Their rendering pipeline is wanting in some ways, and that'll only be sorted through painful experience. Console emulators (E.g. Yuzu/Ryujinx) are in the same vein as Proton, so if you want a health check on what problems something like Proton for Apple would face, you can see what problems the emulators are currently dealing with//hacking around.

I don’t know. They’ve always touted games from time to time. The key limitation is that AAA games are expensive, game developers don’t have experience targeting Mac, and until Apple starts shelling out serious money to seed that market, it’s probably not going to happen. Look at what happened with Google and their failed strategy (there was a lot else wrong, but title availability and porting difficulty was definitely an issue)

They finally have a gaming strategy that makes sense.

Now that iOS devices and Macs have the same CPU cores, GPU cores, and APIs, targeting Apple hardware gives you a much bigger user base that you can target than you've ever seen previously with Intel/Power PC based Macs.

This is the same failed strategy that Microsoft tried with windows 10 uwp. The control schemes and form factors matter so developers don’t get much benefit.

Windows PCs can have such a wide range of hardware that this seems like an attempted software solution to a pretty difficult to solve problem. Compare that to Apple, where there isn't such a disparity between two separate computers, and it makes more sense to try to target MacOS under that logic than it does.

Whether it will play out different is for consumers and developers to decide.

Wow, this is a really good point I haven’t thought of before.

And the thing is, it could work. Because that’s pretty much the main relative advantage consoles have over gaming PCs - extremely narrow range of hardware, which allows consoles squeeze out better performance than a gaming PC with comparable hardware specs (of course the PC will likely be cheaper, but that’s beside the point) on most games, because those games were extremely heavily optimized for that hardware.

MacOS, under that logic, would come through somewhere in-between those two extremes - less disparity than between two random PCs, but more disparity than between two models in a single given console generation.

Even funnier, I just realized that macOS major update support for any given mac is 6 years (and 10 for security updates iirc, but don’t quote me on that), which is about as many years as a single console generation lasts.

It would be interesting to see games primarily designed for consoles get ported to Mac before they work on the Windows port, since the architecture is very similar between the M series APUs and the console APUs :)

> MacOS, under that logic, would come through somewhere in-between those two extremes - less disparity than between two random PCs, but more disparity than between two models in a single given console generation.

Apple is sitting on a gaming console that could be a hit, and is doing literally nothing with it: Apple TV.

Apple also hasn’t fumbled with their platform unification nearly as much as Microsoft did with their attempts. It hasn’t been perfect, Catalyst most notably has warts, but it’s a far sight better than the situation with Windows was where the new platform (UWP and ancestors/cousins) were rebooted multiple times and perpetually half-baked. AppKit and UIKit are both solid, mature, and capable in ways that even a decade after Windows 8, Windows App SDK (what UWP became) isn’t.

Developers already target multiple control schemes (keyboard/mouse, gamepad) and device sizes.

iOS gaming brings in such a large revenue share that it's hard for a business that cares about the bottom line to ignore.

> Mobile games generated $89.6 billion revenue in 2021 across both platforms... iOS was responsible for $52.3 billion of that revenue

https://www.businessofapps.com/data/mobile-games-revenue/

Now that targeting the Mac and targeting iOS devices is largely the same thing, it's much more likely to make financial sense to target Macs than ever before.

For mobile games, sure. For desktop titles, nothing has really changed. MacOS runs the same APIs, Apple still refuses to make compatibility promises, and older titles are perpetually phased-out of support unless they are constantly updated. Apple-native titles will definitely get ported to Mac, but what is the renewed incentive to port PC, Xbox or Playstation games to Mac?

> nothing has really changed

> Resident Evil Village on iOS Is Legit. I Forgot I Wasn't Playing on a Console

https://www.cnet.com/tech/gaming/resident-evil-village-on-io...

And here's the Alien Isolation announcement from two years earlier: https://www.macrumors.com/2021/12/17/alien-isolation-game-re...

And the Tomb Raider 2 release date announcement from 2 years before that: https://forums.macrumors.com/threads/shadow-of-the-tomb-raid...

Every time, Apple pays a big studio to port their AAA title and change things forever. Every time, nothing changes.

You don't have to beg or bribe studios to get them to target a market with 90 Billion dollars in annual revenue.

Targeting that market gives you a Mac version for very little additional cost.

...which is back around to my original point:

> For mobile games, sure. For desktop titles, nothing has really changed.

So I will ask again: what is the renewed incentive to port to Mac?

> nothing has really changed

The economics have totally changed. Targeting iOS with it's massive revenue share gets you almost all the code you need to target a Mac. All you need to do is add support for a keyboard/mouse UI.

It's the exact same strategy Microsoft used with XBox.

If you're already targeting Windows, targeting XBox required less additional effort (add a gamepad UI to your Windows game) than targeting something like the PlayStation would.

Some iOS devices are now performance-competitive with desktops, and they support bluetooth game controllers (so you don't need (much) of a touch-specific UI)

So you port an existing AAA console title to iOS for the huge user base, and get MacOS for relatively little engineering cost.

I assume 90%+ of that is gambling or other mechanics where the spending of addicted “whales” makes up most of the revenue.

> They finally have a gaming strategy that makes sense.

Do they?

> Now that iOS devices and Macs have the same CPU cores, GPU cores, and APIs, targeting Apple hardware gives you a much bigger user base

If you lump mobile phones into a single category with desktop computers and laptops, you don't have a strategy. You have Apple Arcade with mobile-only games pretending they are fit for a big screen.

Gaming strategy is what Microsoft had in the 90s early 2000s when they were going full steam ahead with DirectX.

> You have Apple Arcade with mobile-only games pretending they are fit for a big screen.

Full-blown Resident Evil Village is available on iOS with graphics that are very console-tier.

Not saying that this and a few other counterexamples somehow mean that their gaming strategy will succeed. But I wouldn’t say that including iOS in their gaming strategy would fail due to iOS hardware limiting them to mobile-tier current Apple Arcade type of games.

> Full-blown Resident Evil Village is available on iOS with graphics that are very console-tier.

You know the difference? Consoles are usually attached to big-ass TVs. So whatever graphics iPhones may have games like Resident Evil will be few and far in between, and more of a curiosity.

We already had "console-level graphics" with whatever was the fantasy demo it was 10 years back. Nothing materialized then. Nothing will materialize now.

> Consoles are usually attached to big-ass TVs.

So plug the phone into the big ass TV when you want to do that? Or just play on an iPad?

Ah yes, because that's what I will gladly do: plug in my perosnal communicationd evice into a TV to play... what exactly?

> Or just play on an iPad?

"just". I don't have an iPad. I do have a gaming console and a TV, a gaming computer (PC), and a work laptop (Macbook).

> You know the difference? Consoles are usually attached to big-ass TVs.

You can just connect your iPhone/iPad to your TV using a USB-C to HDMI cable, and all is good. I bought a random USB-C to HDMI cable off Amazon, and it simply worked. No fiddling with settings, no setup, I just connect the cable and everything simply works.

The support for gamepads on iOS/iPadOS is great too, tried it with first-party PS4, PS5, and Switch (Switch Pro controller) gamepads. Worked without any issues, the whole pairing process took less than a minute (and that minute included googling how to do it).

I was moving across the country recently, so for a period of time I didn't have housing and had to stay at my friend's place (which he generously allowed me stay at while he was out of town for a week). I just connected my iPad to his TV with a cable, paired my PS4 gamepad with the iPad, and made quite a bit of progress in Vampire Survivors this way.

I can't imagine many people connecting their phones (with all the usual distractions of notifications, chats, emails etc) to an HDMI cable randomly hanging from the back of their TV to play a game.

> I was moving across the country recently, so for a period of time I didn't have housing and had to stay at my friend's place

While there are use cases for this here and there, does it work as a gaming strategy for a company?

> I can't imagine many people connecting their phones (with all the usual distractions of notifications, chats, emails etc)

Good news, by default iOS suppresses all notifications when connected to an external display :)

> While there are use cases for this here and there, does it work as a gaming strategy for a company?

For someone like me? Agreed, it doesn’t really make sense. For someone whose only potential gaming device is a smartphone? I don’t know, perchance it could act as a gateway to upselling a macbook to users or just as a side act to macOS gaming. But that’s a very shaky guess on my part.

So rich enough to buy an expensive "Pro" last generation phone from Apple but not rich enough to also purchase a dedicated console to connect to his TV so he can play without draining the battery and monopolizing his primary communication device. The Venn diagram of this must be quite funny. You Apple guys are delusional, mobile gaming doesn't make a lot of sense, in particular for the people who have enough money to buy Apple hardware. Apple sucks at gaming and forever will because for it to be competitive they would need to sell performant hardware at an actually competitive pricing. But since Apple like to pretend their parts are made of gold, this is not going to happen anytime soon...

> Gaming strategy is what Microsoft had in the 90s early 2000s when they were going full steam ahead with DirectX

Apple is literally repeating Microsoft's strategy for breaking the XBox into the console market.

Targeting the high volume platform (Windows/iOS) gives you code that is nearly identical to the code you need on the lower volume platform (XBox/Mac). All you need to do is support an additional user interaction method (Keyboard Mouse/Gamepad/Touch).

> Apple is literally repeating Microsoft's strategy for breaking the XBox into the console market.

You know the difference between XBox and iPhones? One is a dedicated gaming platform plugged into a big TV. The other is a personal communication device with a small screen.

> All you need to do is support an additional user interaction method (Keyboard Mouse/Gamepad/Touch).

Ah yes. "Just" support a different interacton model. Nevermind the plethore of issues with that and the differences between big and small screens.

Because the Vision Pro has no compelling use case. Games is what you do when you don’t know what to do.

Or anime chatrooms.

You're telling me "ignore my son's birthday party to watch Youtube videos on my face" isn't a compelling use case?? Staring awkwardly at my wife and saying "Yes sweetie I am definitely giving you my full attention" while I definitely am not is out too?

Face computers have a lot of convincing to do before they are adopted widely.

> You're telling me "ignore my son's birthday party to watch Youtube videos on my face" isn't a compelling use case??

> Staring awkwardly at my wife and saying "Yes sweetie I am definitely giving you my full attention" while I definitely am not is out too?

You just described two scenarios that are already very common in the present day, if you swap the headset for a smartphone in them.

Not saying it is a good thing at all, it isn’t. But given how many people currently don’t mind doing it with their phones (where their actions are on a clear display), I think significantly more people would be into those use cases with AR headsets (and in the future, glasses). It is literally the same thing as with phones, but more concealed/private. So it would capture both the existing group doing it with phones and plenty of new people.

Note: I am aware that Vision Pro isn’t intended to be worn outside or for prolonged periods of time due to battery and just general headset thing, so no, I am not trying to claim it is going to replace smartphones for people. I was talking more about where this is headed, which will only be possible once the technology gets us to the point of a “regular looking glasses” type of a form factor (or something similarly unobstructive).

Very true, but I believe the public's acceptance of something is really the main test. Google Glass was intended for daily wear by everyone but the public absolutely hated them. Cell phones are accepted, mostly because you can usually tell if someone is recording you (compared to Glass) but personally I have thought AR glasses would be awesome.

I could have an AR Pokémon pet, like Pokémon Go...but good...

I'm curious -- is there an ~objective comparison of how modern Macs (M1, M2, M3) compare to PS5, Xbox X/S, PS4, and Xbox One? And how it differs between passive cooling (MacBook Air) and active (Pro, Mini)?

In terms of GPU and CPU and memory?

I mean, could a M1 MacBook Air or Pro run a hypothetical native version of GTA V as well as the PS4 did? I'm assuming it can't match a PS5, but how close could it get?

It’s really hard to compare because to exploit the Apple hardware to the fullest you have to rearchitect your code to run on Metal. That has been done for ~0% of the game code out there. So real world comparisons are rare.

By nominal TFLOPS, the GPUs in the M1 Max, M2 Max, and M3 Max are in the same neighborhood as PS5 and Xbox Series X. On the CPU side, even the base M1 is faster than any console.

Apple pricing model creates artificial scarcity with memory and storage, which makes the base models poorly suited for gaming. Consoles ship with 16 GB memory and 1 TB SSD, while reaching the same numbers is still a $600 upgrade to the base M2/M3 Macs.

I recently tried arm64 Factorio on an M1 Mac, and it's smooth as silk. It also loads within a couple of seconds even with dozens of Bob+Angels mods.

The Mac seems to be only slightly warm while playing. I haven't yet played to the point of building enormous 4 lane railway squares, I'm intrigued to see just how much the M1 can handle.

Great for entertainment, absolutely terrible for my productivity!

Steve Jobs started out working on arcade games with Woz. Despite that, I got the impression he didn't care at all for video games.

From what I've heard, the person who was originally in charge of the iOS App Store completely dismissed videogames altogether. He believed music was the only art form worthy of Apple. Rumor has it he actively sabotaged the effort to make Apple TV into a market-competitive game console. When he was moved to the iTunes team, iOS suddenly appreciated games.

Between the two of them, Apple has been actively steered into not being a platform for games. Maybe that's finally (very slowly) changing.

Honestly, given the crappy direction that windows is taking, I'm ready to abandon it the minute apple can run games seamlessly. And I doubt I'm the only one.

> "Apple relies on Qualcomm for its modems, although in 2019, the two companies settled a two-year legal battle over intellectual property. Soon after, Apple bought the majority of Intel’s 5G modem business for $1 billion, in a likely move to develop its own cellular modem. That hasn’t happened yet, and in September, Apple signed on with Qualcomm to supply its modems through 2026."

After Qualcomm proved that Apple was forwarding restricted technical documents to Intel, it seems likely that Qualcomm will take a very close view of any Apple modem, and vigorously enforce any patents that Apple is found to infringe.

Bearing this in mind, why not second-source Mediatek 5g modems in a few models? An in-house effort is bound to keep the lawyers very busy for years to come.

https://www.nytimes.com/2018/09/25/technology/qualcomm-apple...

> “Qualcomm will take a very close view of any Apple modem”

Doubtful.

Apple & Qualcomm settled such dispute.

https://www.apple.com/newsroom/2019/04/qualcomm-and-apple-ag...

Qualcomm would not look? Come on... Qualcomm is a lawyer IP machine that happens to make cell chips.

From the link in my previous post:

> ” The companies also have reached a six-year license agreement, effective as of April 1, 2019, including a two-year option to extend, and a multiyear chipset supply agreement”

And Apple has already extended that agreement.

https://www.qualcomm.com/news/releases/2023/09/qualcomm-anno...

Apple has paid to have all rights, and the companies have publicly stated they have settled.

As such, there’s nothing for a Qualcomm lawyer to audit anymore.

This article appears to have been written by someone who heard about Apple's chips and, since it was new to them, decided it was actually new.

Respect to Apple, but internal ASIC teams is nothing new in a lot of the tech industry, and I don't mean Google's TMU stuff or the Annapurna team at AWS. Those guys were following a long trend of ASIC teams in companies - e.g., ALL of the networking industry except Arista has been running this way for the last three decades.

I mean, after the 68k era, the _entire Unix workstation market_ - IBM, HP, DEC, SGI, etc. - was entirely built around in-house ASIC teams building their own CPUs.

I find it exciting, but what it really is, at the end of the day, is verticalization so that they don't have _as deep_ of a margin stack on their CPUs. Microsoft and Sony have gotten a series of custom CPUs and GPUs from AMD, Nvidia, etc. over the last twenty years without owning the whole stack or losing the ability to apply optimizations. That's been common forever.

Apple is doing an amazing job. It started with acquiring the PA Semi team and has grown from there, they've been amazing and consistent about it and it's awesome execution and _sticking to the investment profile required_, but .. eh, the article just seems utterly unfamiliar with any broader context.

>Respect to Apple, but internal ASIC teams is nothing new in a lot of the tech industry

It's not new indeed, but Apple was the first to turn custom in house chips into a major selling point. For example Samsung was vertically integrated for a long time, with their in-house Exynos chips and modems, yet they couldn't capitalize with that in their phones, as they often fell behind Qualcomm in CPU-GPU performance, modem reception and battery life.

Being vertically integrated is not enough if those parts you're making in house are mediocre and don't enhance the final product.

No, they aren’t. Literally any of the examples I give did that. Not in consumer, but elsewhere.

Another thing that everyone has been doing for decades: “We built what we call the unified memory architecture that is scalable across products”

> But there’s a lot more to the biggest U.S. company by market cap than just gadgets.

What a dumb statement. Those "gadgets" would not exist without those chips (and a bunch more besides).

Also, if the word "gadget" is appropriate, it doesn't apply to the Mac (admittedly Apple's smallest segment). But I think it no longer applies to the phone either.

Something I never quite realized – Johny Srouji's been at Apple for a while, even it looks like before the PA Semi acquisition. Keeping up with the chip industry is one thing; keeping up with the chip industry in an organization like Apple is another.

As a source of both continuity, and in terms of influence on Apple's long-term vision as a company, that's amazing.

Reply on news.ycombinator.com