kzeise.com

If we look at the GeekBench scores of the 6s in detail, we see clearly see where the improvements come from. Memory copy is roughly 45% faster, this indicates that Apple has done a good job improving the interface to the main memory. Taken together with the now 3MB of level 2 cache, it is no surprise that the LUA scores have made a huge jump, with the interpreter causing a lot less misses. The AES scores see the lowest improvement, since it is almost completely hardware accelerated, and so only profits from the increased frequency. It is also noteworthy to see the improvements in SHA1 and some of the filters; these indicate that Apple must have improved the pipelining of a few commands, reducing the stalls from waiting for a result to arrive. I believe these pipeline stalls are also behind Apples push to have apps stored in an intermediate form on the App Store, this would allow Apple to change the processor behavior even if it meant that code optimized for an older processor would run slower.

There is one number that leaves me scratching my head from the Apple Event: the 70% performance increase for the A9 processor. This is an astonishing jump in performance given that Apple was already using a 20nm process for the A8. It will be interesting to see how this increase is achieved. Possible avenues are more cores, higher frequencies enabled by better controlling leakages, and better optimized logics. Given that Apple has already caught up to Intel in performance per cycle and core, I would have assumed that Apple would have continued with the slower pace set by the A8. 

With slowing sales for the iPad, while the iPhone is going from strength to strength, Neil Cybart wonders if it has already plateaued, if it turns out to be a consumption device for the web and Netflix, and not a computer to replace the Mac. 

I actually believe that the iPad is loosing sales because the iPhone is supplanting it as the better consumption device, that the 6 plus is the device those iPad mini owners really wanted, and that the iPad has not yet improved to the point where you can do more complicated tasks easily.

While Apple will support running two apps side by side with iOS 9, it still has work to do to replace the collaboration that is enabled in a crude manner by the file system on PCs. Add to this that the App Store is not very supportive to financially support quality applications, not to mention the risk of censorship by Apple, and we are not seeing many new, great apps: iPhoto discontinued, and I do not remember any great new apps since Pixelmator, OmniGraffle, Paper. 

On the hardware side, I believe we slowly realize that the iPad is portable, but not mobile computing, and that it is therefore reasonable to trade a larger surface for better productivity, and accept a bit of extra bulk as the trade off. The trouble is that the threshold for something you can comfortably hold in your hands for longer periods is roughly 400g, maybe 500g. Once we are able to buy an iPad with whole day battery life and a 12″ screen area for this weight, it will replace most PCs and Macs in use.

The iPhone has a neat feature to reply to an incoming call with a prerecorded text. I find the standard texts pretty useless, and have them replaced with: Sorry, I’m busy. I’ll call you back in 30 minutes, plus variants for 2 hours, and tonight. I choose those because someone calling me typically does not want to know why I cannot respond, but when she can expect the call back. And I can explain the why in the phone call later, with more empathy.

David Smith feels the activity rings are too hard to read. I agree, they are hard to discern between almost closed and closed. But I feel a better way to show this would be to have the last quarter of the ring fade out, and then restore it to normal again once closed. This is because we are much better at seeing differences than absolute values, and with David’s suggestion you would be hard pressed to see the difference between three almost closed and three closed rings at a glance.

In order to find out how one would value Apple shares, one starts out by looking at the market potential of its products, iPhone, Watch, iPad and Macs. 

The iPhone has the widest appeal, and it is safe to assume that Apple will continue to be able to corner the high end market. As a rough estimate, I see the following numbers:

As the world gets richer, this number will slowly rise, and could reach a billion within two decades. 

For the iPad, I believe the market to be a bit smaller, since the 6+ size is good enough to fulfill most consumption roles, so let’s make that 450m users. The Watch has great potential, and as such I see two thirds of iPhone users getting one eventually, say 400m users. The Mac is more limited, as it appeals to professionals. Let us say that the market is 150m users. 

To get to the addressable market, we need to know how often users will replace their units. We have seen from the last iPhone upgrade that the age of annual performance doublings is over, and knowing that only 20% of iPhone users have upgraded in the last half year, I estimate that we will see a significant lengthening of upgrade cycles: 4 years for the iPhone and Mac, 6 years for the Watch and iPad. This leads me to the following estimates:

With 5.82b outstanding shares that would be an EPS of $11, with some extra profit from services. While there is a risk, that as the technology matures, we will see a longer replacement cycle, longer lifetimes will also increase the lure of purchasing a quality product. Even though everybody knows what people value with Apple, namely superior quality, design, customer service, and especially the experience of using the products, there has been nobody able to compete. So the chance that Apple will be replaced as the quality leader is relatively low, with Xiaomi being the greatest threat.

With the 2015 MacBook, Apple has provided us with a lightweight solution for people with average computing needs, and a preference to use a keyboard to get their tasks done. I have not seen yet a breakdown where the average 5W power draw goes, but we have three main components that draw the bulk of power:

Screen needs to power the backlight, and the individual pixels in the LCD. Most of the consumption is with the backlight, typically dominating the entire power budget at the highest setting, under low loads.

DRAM Memory can draw a few watts of power when actively accessed, while the self refresh is less than 10mW per GByte. More memory increases power consumption mainly because chips not accessed in a transaction still increase parasitic power draws on the connections. The best strategy to reduce power consumption is to not access the memory, and this means software optimized to utilize the RAM caches. Intel Iris Pro provides a 128 Mbyte extra cache which can also greatly reduce main memory accesses, even though currently its own consumption eats up any power benefits.

CPU/GPU are the classical consumers, but have now become so powerful that they tend to sleep almost all the time. When we actually need to do some processing, it is worth remembering that power consumption is proportional to f U^2, the frequency times the supply voltage squared. Since with a higher frequency we also need a higher voltage so that the transistors can switch fast enough, we can compute more with a given charge by running slower, up to a point.

So when Marco Arment wants more battery life, we can actually configure a computer that can run long on batteries, and fast when plugged in. We can set up the latest processors to disable Turbo Boost, even reduce its TDP to reduce cooling requirements, so that we do not need waste energy spinning a fan. The external GPU can be completely disabled while on battery. One can change the system scheduler to reduce the number of active cores when it needs to boost a single threaded load. The envelope to throttle down the available power to increase battery life into the 12 hour plus range is there, it is up to Apple to actually implement it.

Marco’s recent experience using a 6 Plus reminds me what still annoys me about its smaller sibling.

The battery is too small when traveling. Most of the time the battery life is just fine, but whenever I am traveling, and using the phone very intensely to check maps, track my steps, get information, and take pictures, I have to restrict my usage to make it through the day. Part of the problem is that the automatic dimming sets the screen brightness very high outdoors, but otherwise we just need a larger battery.

The case is too slippery. Mainly because you no longer can properly grip the phone, it now becomes mandatory to use a case to get enough friction to have the phone not slip from your hands.

The keyboard is too large for one handed use. The 6 works remarkably well one-handed, provided the back is not slippery thanks to a case, apart from the problems with typing. Here we would really need a narrower keyboard, the size of the 5 keyboard shifted to the side of your thumb. This is a hard design problem, because then the phone would ideally need to figure out how you grip your phone to present you with the correct layout. But maybe we could detect a sideways swipe on the space bar to quickly move the keyboard around?

Example of keyboard shifted to the right

In general I would like to see the next iPhone prioritize whole day battery life while traveling above extra thinness.

The 2015 MacBook looks great: a very light computer, with a Retina screen. It is essentially an improved version of what has made the MacBook Air great, following the design direction provided by the iPad. That it lacks some of the ports we have come to expect is just the natural evolution coming from ever improved wireless connectivity. And none of these ports will be missed in a few years time:

• The SD card was great to store photos, but its time has come. With flash memory now cheap enough for tens of thousands of photos, it is high time dedicated cameras get just a large internal storage, and provide wireless access to transfer photos to your computer and phone in the background.

• Almost nobody carries anything else with them to connect to their computer when on the road. And for those items at your workplace a dock is a good solution, with only one cable to connect to your computer.

When you are using an external monitor, it is the natural place to double as your dock. And on the road you will probably want to have a combined dock and power supply that can charge not only your computer, but also your phone and tablet. The electronics needed for this can be made compact enough to not add much bulk compared to a pure power supply.

The unique experience provided by the Apple Watch is an always present user interface, crammed in a very small space, which makes interacting difficult. Given that it will be difficult to actively choose between multiple apps, we will see proximity as a tool to select the expected app for us: shopping list at the grocery, car keys at the car, thermostat at given points in the house.

We will see new apps that only interact with specific physical locations. Tools will appear that will replace their controls with the watch and a Bluetooth connection. So instead of turning a knob you will touch your watch, tap on the function you want to modify, and turn the crown. It is essentially what location based glances are for, and they can be better than the physical interface since you can modify things also when a couple of meters away. But it also needs quick access to control the nearest thing, maybe as a complication on the watch face.

This is similar to what Ben Thompson is saying on the Watch adoptation: The watch will be a great interface to all the internet of things items that people are dreaming about, but it will also enable your environment to better track you. So a room could know who is inside, and set the temperature automatically based on their preferences.

Given the huge amount of data the watch will collect (or enable others to collect), it makes it critical that this data is protected, well separated between the different hats we wear during the week, and not shared without our active consent. It underlines why Apple is best placed to make this a success, and why it is so critically important for Apple to keep that trust that Apple will not sell data and will not make devices that can be hacked. Especially with Apple Pay and other authentication solutions a hack could bankrupt you. So Apple cannot tolerate that the NSA, never mind criminals, can hack into your watch storing your car keys, setting them on a collision course.

In addition I believe that the iPhone is critically important to bridge the adaptation gap, to provide in the iPhone a fallback that works reasonably well to bridge the time until enough people are owning one.