Also, batteries will degrade faster over time when they start to degrade, because they need more frequent charging. Their internal resistance increase and that promotes heat buildup during fast charging/discharging, another thing that promotes degradation. Slow charge/discharge cycles also help with heat management.
Having a Windows 11 corporate laptop with a domain/Entra login, I actually trust it more than a home Windows 11 with a Microsoft account. Because if I lock myself out, I have a contact (corporate support) that is actually interested in helping me recover everything. With a Microsoft account it's a mess. I had so many problems with Microsoft accounts that I lost count of how many I have, and most are broken in some way, because of different issues and different service integrations over time. The Skype account is now useless. I never recovered my paid Minecraft account after one event. With a machine with a local account, now I have to be very careful on what I click related to MS accounts, because trying to solve various issues with Teams, I managed to get the local account linked with that MS account. I spent hours trying to recover a different account after I randomly filled one nagging question about birth date - who wants to give the real birth date to Microsoft - and then I got locked out because I said was underage :). So yes, one of the big issues is the push to have a linked OS account where you have to rely on MS support to solve your issues, otherwise you basically get locked out of your machine and other things you paid for.
Also, domain policies offer more control over the corporate PCs (this is how some of the MS spying is shut off on corporate PCs; it's debatable if the corporate spying added by other domain policies is an improvement).
I have to agree, I've also suffered account problems. I was locked out from an email address I used for 20 years. It refuses to take my password which is still valid. I've changed phone number since 20 years ago so can't use that and the security questions were nonsense as I was a teenager. Originally my account never had phone number, they insisted I add it when they integrated my Skype account perhaps. So I didn't expect access to that phone number to be a strong ongoing requirement.
With mirrorless cameras the focus switched from specialized sensors to on-CMOS contiuous exposure sensors, so movement is easy to detect. At this point the cameras have specialized AI hardware to run the models, and they also accept user input (on R5 MkII you can register up to ten people to prioritize focus on[1]). The focusing options are now very complex[2][3], and combined with lots of customization options on the camera's buttons you can have very specialized/personalized setups for different types of photography.
Sure, as I said in the first paragraph, AF is these days very impressive thanks to the large amount of data available (but of course this would have been too much data back in the day, when there wasn't nearly enough CPU power to process it fast enough). I wanted to give more historical context for how AF worked before fancy AI.
The AF settings, except those related to face/object recognition, haven't actually changed that much since the 7D Mk II days. The preset system is more general now and allows you to store and recall all AF settings rather than just the three tracking-related variables. The high-end DSLRs used to have six cases for different types of sports that you could modify but not rename.
The usual problems with this approach are 1) the roof is not big enough for solar panels to power multiple chargers and 2) the power is generated when people are not at home (and probably using their car to go to work).
Then the obvious solution is to put solar panels over parking spots at work. Doesn't need to be that much power, 1-2kW is enough if the car sits there all day (for typical distances people commute with cars)
Yeah, it's basically dying and living on old fame. We had to buy a license a few years ago for a customer who needed support for some things that PDFBox did not support, Of course it's not just a license, you have to buy multiple licenses for production, development and so on. It was okay until we hit a bug, iText could not read some form fields correctly and was basically changing the PDF contents on save. We opened a support ticket with all the details, sample files and code to reproduce. The ticket stayed open for a year. After a year they asked us to pay for more support. We showed them the open ticket and never heard from them again.
The problem is that a cable saying that it's something does not mean that it actually conforms to the standard and can deliver that speed/power. And proper testing requires complex hardware.
That's true of every cable that has ever existed, including basic electrical wiring, and so feels like a separate problem than merely identifying what kind of cable a cable even intends/claims to be in a world where they all have the same shape connector: don't let the perfect be the enemy of the good.
It will be faster only for code that uses/is optimized for that specific extension. And the examples you give are not really correct.
If you add a supercharger you will get more power, but if the car's transmission is not upgraded, you might just get some broken gears and shafts.
If you add more solar panels to your roof, you might exceed the inverter power, and the panels will not bring benefits.
It's true that you will benefit from the changes above, but not just by themselves - something else needs to change so you can benefit. And in the case of the M4 and these extensions, the software needs do be changed and also to have an use case for these extensions.
That is an example of the kind of hardware and software synergy that gives good performance on Apple Silicon for apps in iOS and Mac OS. Apple execs have given interviews where they talk about this kind of thing. They look at code in the OS and in their application libraries that can benefit from hardware optimization and they build in hardware support for it to improve performance overall. This helps all kinds of app running on the OS and using the standard libraries.
Are you implying there are no use cases for matrix multiply?
In any case, the two main deep learning packages have already been updated so for the place this change was almost certainly targeted for, your complaint is answered. I'm just stunned that anyone would complain about hardware matrix multiplication? I've wondered why that hasn't been ubiquitous for the past 20 years.
Everyone should make that improvement in their hardware. Everyone should get rid of code implementing matrix mult and make the hardware call instead. It's common sense. Not to put too fine a point on it, but your complaint assumes that GeekBench is based on code that has implemented all those changes.
> Are you implying there are no use cases for matrix multiply?
The whole point is that these highly specialized scenarios are only featured in very specialized usecases, and don't reflect in overall performance.
We've been dealing with the regular release of specialized processor operations for a couple of decades. This story is not new. You see cherry-picked microbenchmarks used to plot impressive bar charts, immediately followed by the realization that a) in general this sort of operator is rarely invoked with enough frequency to be noticeable, b) you need to build code with specialized flags to get software to actually leverage this feature, c) even then it's only noticeable in very specialized workloads that already run on the background.
I still recall when fused multiply-add was such a game changer because everyone used polynomials and these operations would triple performance. Not the case.
And more to the point, do you believe that matrix multiplication is a breakthrough discovery that is only now surfacing? Computers were being designed around matrix operations way before they were even considered to be in a household.
I'm not complaining, I'm just saying that the higher numbers of that benchmark result do not translate directly to better performance for all software you run. Deep learning as it is right now is probably the main application that benefits from this extension (and probably the reason why it was added in hardware at this point in time).
Well you're really just describing benchmarks- if the benchmark doesn't represent your standard workflow then it probably isn't a good reference for you. But Geekbench includes a bunch of components based on real-world applications like file compression, web browsing, and PDF rendering. So it probably isn't perfect, but it's likely that the M4 will feel a bit faster in regular use compared to an older generation MacBook Pro.
The article says "100 times greater than TDK’s current battery in mass production" but they are not referring to the current LiIon/LiPo batteries, but the current _solid state_ battery. The capacity per liter of the new solid state battery is less than 2x of the current phone batteries (1000Wh/liter vs 5-700Wh/liter for LiPo). So no, you cannot replace one phone battery with a coin cell with the same battery life.
That seems to be a modified bike, with a head tube extension plus lots of random spacers, and lots of accessories. It's probably used as a very comfortable road/gravel bike.
