mr. h

It's a shame you didn't get this article checked by someone who actually knows anything about power electronics. It's really quite confused in multiple ways.

Firstly, it refers to GaN "chips" implying that it's being used here for integrated electronics (i.e. a single die containing multiple transistors, like a microprocessor), but that's not what's happening. These chargers are using established switched-mode power converter topologies and replacing silicon (Si) MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors, a Si die consisting of a single transistor, that is able to handle lots of voltage and current, relative to a microprocessor transistor) with GaN HEMTs (High-Electron-Mobility Transistors).

Secondly, it refers to voltage being "conducted". Power transistors do not "conduct" voltage. In power converters like these, the power transistors operate in a "switched mode", being either "off", or "fully" on (where they behave like a very small resistance in the tens of milliohms range). When "on", they conduct current, and when "off" they block voltage.

When conducting, the power transistor dissipates some power (referred to as "conduction loss"), this being equal to the current squared multiplied by the resistance. So, lower resistance = lower losses. However, when making the transistor, there is a tradeoff between how many volts the transistor can block when off, and how low its resistance is. The more volts you need to block (in the case of these power converters, the transistors are usually rated at around 600 V to 650 V), the higher the resistance. The wide bandgap of GaN gives a better tradeoff between voltage blocking capability and on-resistance, compared to Si. Additionally, when transitioning from "on" to "off" or visa-versa, there is a brief period of very high power dissipation in the transistor (referred to as "switching loss"). Crudely speaking, the faster a transistor can perform these transitions, the lower the switching loss will be. GaN power transistors are much faster than Si devices and the switching losses are much smaller.

At a very basic level, switching power converters operate by taking small packets of energy from the source (in this case a mains/grid connection), storing that energy temporarily, and then outputting that energy to the load. The energy is stored internally in inductors and capacitors, and routed/transferred with power transistors and transformers. How often the packets of energy are routed around is referred to as the switching frequency. The higher the switching frequency, the smaller the energy storage components and transformers can be. However, as mentioned previously, each time a transistor transitions from "on" to "off" or visa versa, it wastes some energy as heat. So the higher the switching frequency, the higher the losses. If in each transition you lose less heat, you can afford to increase the switching frequency and therefore make the inductors, capacitors, and transformers in your power converter smaller.

In summary, the key benefits of GaN over Si, when it comes to power transistors, are:

A better tradeoff between on-state resistance and blocking voltage (lower conduction loss)
Much faster switching transitions leading to lower switching losses, allowing a higher switching frequency.

Now, both GaN and Si transistors have an upper limit of temperature that they can safely operate at, typically 150 C or 175 C. Any power dissipation in the transistor is converted to heat, which must be extracted by a cooling system to keep the transistor below the maximum allowed limit. If you've less heat to extract (due to lower losses), you can make the cooling system smaller. So, there's two ways that the power converter can be made smaller:

1. Get the losses down so the cooling system can shrink
2. Increase the switching frequency to make the inductors, capacitors, and transformers smaller

When compared to power converters using Si transistors, a GaN-based one will be using a mixture of these two approaches.

GaN transistors are also smaller than Si ones, but the impact of that is fairly minimal relative to the other effects outlined above.

To learn more, you can get a free book on power electronics here (registration is required, but free). For the avoidance of doubt, I am not Barry Williams.

pixelwash said:

The earpods Apple ships with current phones are Lightning, and analog

No they are not analog. There is a digital to analog converter and headphone amplifier inside their lightning plug:

http://appleinsider.com/articles/16/09/20/teardown-finds-dac-chips-in-apples-lightning-earpods-lightning-to-35mm-adapter-for-iphone-7/amp/

riclf said:

What I'd like to ask the great minds at AppleInsider and their very smart readership is WHY, if I use a new Sandisk Extreme Portable SSD https://www.sandisk.com/home/ssd/extreme-portable-ssd configured with a USB 3.1 Gen 2 Type-C connector, connected to a new Macbook Pro's Thunderbolt 3 port, do I ONLY get 550MB/s (4Gbps) instead of 10Gbps (1250MB/s) ? Seems like I'm chugging at half speed. What am I missing here?

Is this a trick question? The page you linked to states over and over again that the read speeds are up to 550 MB/s.

bobolicious said:

...and so 'we' agree to our favourite fruit company XX.X release EULA, in context of the represented current business model and in light of the Patriot Act...? I'll ask again where is the off switch for Photos image tagging ?

if you mean location tagging, it would be:

settings->privacy->location services->photos->never

You're welcome.

baconstang said:

chasm said:

Here’s the problem: most third-party parts are probably fine and Apple should have no problem with them.

BUT

It is possible for third-party parts to compromise security. A replaced fingerprint reader, just as a mild example, could not only send your biometrics elsewhere, but could also conceivably compromise the Secure Enclave.

So I totally understand Apple’s hostile attitude to using third-party parts — it cannot be sure which are fine and which aren’t, and should probably start a certification program for third-party parts so that the good ones don’t trigger this issue.

THUS

I doubt it is realistic for Apple to maintain this “you have to use our parts or else” attitude, even if their motives are as pure as the driven snow. Legislation would probably result to bar Apple from requiring only genuine authorized parts, and we REALLY don’t want that.

Just because you MAY repair items yourself doesn't necessarily mean you CAN or should...

It's a pity that this major misconception of the right to repair movement persists. Right to repair is not primarily about allowing individuals to repair their own stuff, it's about providing access to legitimate parts, components, tools, and documentation.

This allows third parties to provide repair services, so if you don't have the skills to repair something yourself, you can take it to an independent repair shop instead of direct to the manufacturer. This often represents a significant saving, as independent repair shops can for example replace one broken component on a motherboard, instead of just replacing the entire motherboard (as Apple would do).

foregoneconclusion said:

hmlongco said:

foregoneconclusion said:

I don't need access to a gigantic database of previously written text in order to type or write. AI programs do. Turn off the database access and they're useless. 

As pointed out, they've been "trained" on a large corpus of information.

As have you, since birth.

Baloney. The "training" IS the database, i.e., exact duplicates of millions of written documents or millions of pixel base images. That's what the "AI" constantly accesses. The idea that computers are going to retain knowledge without exact copies of the material being stored somewhere is ludicrous.

No. That is wrong. That is not how generative AI works. It's how a lot of people think it works, leading to the kinds of fallacious claims that you are making. Generative AI works in a much, much closer way to how your brain works than you either understand or are willing to admit.

An LLM's training and learning really is similar (not identical) to how you learnt your native language as a baby. No-one could "explain" to you what language is - you were just exposed to a lot of it, and your brain had to form new neural pathways to try to figure it out. Your brain detected that it was hearing the same "kinds" of sounds repeatedly, and that re-inforced certain pathways in your brain. That forms the basic scaffold that enabled your brain to then recognise whole words, and from there, start to figure out how to use those words yourself.

AppleZulu said:

What’s described here looks to me like Apple foiling ‘chop shops’ from being able to swap around parts to sell stolen (or otherwise dubiously sourced) iPads. 

Well yes, I guess it does prevent that. But who in their right mind would chop stolen iPads together, when you can just sell them unchopped? Much less effort.

It tends to be a bad idea to limit the freedom of law-abiding citizens in order to prevent actual or perceived criminality. See for example, attempts to add back doors to encryption standards.

The pathetic simping for Apple on display in this thread is utterly nauseating.

Wesley Hilliard said:

M68000 said:

So, now it’s bad to charge your phone to 100% ?  Lol,  so much different information out there.  It’s hard to know what to believe. 

Simple: ignore randombro69 on a forum and listen to the experts. We've detailed what you should and shouldn't do for iPhone battery health. My recommendation will always be "don't worry about it."

Really. This whole thing is getting silly. You can't beat physics.

And I would understand all the drama if battery replacements weren't readily available and cheap.

Who’s the expert here? You? Because your article is dead wrong. It has long been well established that charging lithium ion batteries beyond 80% accelerates the aging process. It’s not the charging per se; rather just being at that elevated state of charge. That’s why the optimised charging mode attempts to wait until the last moment to bring the battery to 100%: the theory is you will start using the phone soon after it reaches 100%, thus minimising the time the battery spends above 80%. However, in my experience the optimised charging is easily thrown off if the user has an inconsistent routine so I welcome this additional control provided by Apple.

foregoneconclusion said:

mr. h said:

Highly advanced cut-and-paste.

Is a gross over-simplification of what these tools do. In fact, it may not even be an oversimplification, but rather just flat-out wrong. When you are typing, are you performing "highly advanced cut-and-paste"?

I don't need access to a gigantic database of previously written text in order to type or write. AI programs do. Turn off the database access and they're useless. 

No, that is not what an LLM is. The model was trained on a large database, but it does not access that database when generating responses.

gatorguy said:

I get that you really REALLY want to paint a glowing picture of "gosh Apple is doing this for us", but is there any even circumstantial evidence Apple was ready to make everything end-to-end encrypted in a way they could not access any of your data even if they were ordered to? Not as far as I know. It's more of a hope and prayer since otherwise it's not for the betterment of us users. 

All I can say about that is that the whole scheme would be totally pointless if they weren't going to encrypt the photos. Why go to all the effort of designing this enormously complicated system, calculating hashes on-device, doing the CSAM hash-matching in a "blind" way so even the device itself doesn't know if there's been a match, and then going to all the convoluted effort of generating doubly-encrypted "vouchers" and associated "image information", if the photo itself was uploaded to iCloud unencrypted?

Certainly, this system would enable the photos to be uploaded to iCloud encrypted, but I concede that as far as I know, Apple hasn't said that they would do that. It's just that, as I said, the whole scheme seems totally pointless if the photos are uploaded to the server in the clear anyway.

How about Apple just offers a toggle in iCloud photos settings? The two options would be:

1. Photos are CSAM-scanned and encrypted before being uploaded to iCloud.
2. Photos are not CSAM-scanned, but are uploaded to iCloud in the clear. The server then does the CSAM scan.

Would this solution make everyone happier?