There isn't code for when to pass. It's part of the neutral net it learned from its training. While to a human player it would be immediately obvious that passing when behind is the same as giving up, the AI didn't learn this. It just didn't come up in the training set often enough.

A 20 mph speed limit on all roads would save thousands of children's lives. Why are those children's lives worth so much less when the law that would save them would inconvenience you?

There's an incorrect assumption there. The assumption is that the biomass would still exist without the cows. That's not true. It's created to feed the cows. It wouldn't transform to methane another way because it wouldn't exist in the first place.

But auto formatters are always zero exception. Make a readable table, with one item per row and columns that line up, and then the auto-formatter will come pack everything to the line length limit and wrap it where ever. Becomes unreadable.

I think the reason some people need zero-exception strict style guides is that they can't actually write code in any kind of style themselves. Don't even know what the space bar, tab key, and enter key are for. The must have an auto formatter do this for them. It's like riding a bike without training wheels. Balance myself? No way, I can't do that! Any deviation, no matter how much more readable it makes something, must be eliminated because that's what their auto formatter does.

The importance of strict style guides is just a cover for their own lack of skill at being able to write code on their own.

It's a bit misleading to say it beat a 50 year old record. Strassen's method from 1969 has been bested 11 times according to Wikipedia (Sorry, I didn't search for every paper on fast matrix multiplication myself) since then, the latest in 2020.

Strassen can multiply two 2x2 matrices with just 7 multiplications instead of the 8 it takes using the straightforward method taught in high school. But it's not used to multiply 2x2 matrices in practice, because saving that one multiply costs extra additions. And it processes the data in a slower, out of order way. And there are a bunch of extra temporary values to keep track of. So it's slower. Even without vector instructions it's slower.

But it can work in a divide and conquer way too. Think of a 4x4 matrix as 2x2 matrix made of four 2x2 matrices. We can multiply it the normal way by doing 8 multiplies, but each multiply is of 2x2 matrices this time (for a total of 8*8 = 64 scalar multiplies). Strassen's method lets us save one multiply, but this time it's not just one scalar multiply operation, it's a whole 2x2 matrix multiply that gets to be saved.

It's still slower! But eventually with big enough matrices, it gets to be faster. Like 512x512 big at least.

So they found something better than Strassen for small matrices. But it's not actually useful for those sizes. And it's unlikely their method will be either. Now if they had something for large matrices, it might actually be faster than what's used now.

But the sunlight hitting the water also generates heat. So it's not necessarily going to heat the water more to have a panel over it than it would have been directly exposed to the sun. One would think cooler, since at least some portion of the energy from the sun is converted to electricity instead of heat. Rooftop solar panels reduce attic temperatures, so I'm going to have to go with colder, not hotter, because of the panels.

Looks like the next max is an Amlogic T931. I wonder if they will be able to port it to a current phone processor?

Because many of the silicon vendors, especially those with names beginning with Q, keep their chips secret. There's no docs. They give you an old kernel that's massively hacked, has a ton of binary firmware blobs, and probably a half dozen or more bootloaders that are binary only too.

Merely trying to port it to a newer version of the Linux kernel is all but impossible. Do Google's flagship phones run a current kernel? No. Because they can't port you-know-who's BSP to a new kernel.

Trying to port to an entirely new operating system seems all but impossible. The chip would be EOL before you finished! (One notes Fuchsia has previously only been ported to discontinued devices...)

So unless they switch Android to Fuchsia and force every handset maker to go along, I don't see how they'll get the silicon vendors to support it, and without their support, I don't see how they'll keep up with the pace of processor development.

Libraries don't come with keys. You can buy WolfSSL or use OpenSSL non-FIPS version, that's unrelated to knowing how to use it and being stupid enough to copy an example key into the product.

Many consider the FIPS version of OpenSSL to be worse. Once something is certified, it because prohibitively expensive to change it and re-certify. So FIPS-OpenSSL is basically an old version on OpenSSL, with unfixed exploits.

USB-C connectors can run display port as one of the alternate modes. That's the most common way video over a usb-c connector works. I think there's a usb-c mode for HDMI too that no one ever used.

If you want a stripped down distro for a specific purpose, then there are already distributions for that. If your making an embedded application, then use Yocto, Buildroot, or OpenWRT. They are designed to be small. And they are designed around cross building, so one doesn't need to build on the device itself or even the same CPU architecture.

At one point I worked for a silicon vendor, and developed the Linux firmware for a product using one of our chips that was not yet in production. I had to write kernel drivers for things as yet unsupported. Like SGMII for the phy or even the GPIO pins! BTW, this is where the Linux kernel sysfs interfaces to gpios came from. I didn't make a Linux distro from scratch.

There's a huge difference between creating and maintaining your own distro, and only being able to install a binary distro someone else built for you. The former is a waste of duplicated effort doomed to always be inferior to existing distros with more users and man-hours of development, the latter is for someone who isn't much of Linux engineer. If someone lacks the skills to use Yocto or Buildroot on their new hardware, how were they ever going to create a distro from scratch?

I've never understood the desire to maintain a Linux distribution. It's a ton of never-ending work and there are already dozens of different groups doing it.

And I've worked for two companies that did this. One maintained a custom Linux distro for their products and the other was a silicon vendor who did it for their BSPs. To be fair, both started doing this a long time ago before there were as many options. But it was clear this custom distro was acting like a brake to development. They were always behind. Old packages, integration bugs, unfixed CVEs, lacking in packages and libraries, old toolchains, etc. Man hours had to be constantly devoted to keeping them up to date and pull in fixes that were already present in other distros.

One company has since switched to yocto and the other no longer exists.

While it's true current events might be having an effect now, it looks like there's been fixes queued that haven't been acted on since October 2021.

Killing off the disease spreading mosquitoes will mean more people. Historically, more people have not been helpful to ecosystems. I suppose it's been good for corn, wheat, cows, chickens, etc., but for everything else, not so much.

You are taught how to tack in the first lesson. If you weren't, how would you ever get back to the dock? Wait for the wind to change direction?