If they're selling below cost, who's the "we" that should buy them all up? Because if it's US consumers, then the domestic auto companies could fail in the while that's happening. If it's the US government - isn't that kind of what the tariff accomplishes, without having to take ownership of a bunch of inventory?
I think we can confidently call stupid actions stupid actions. Inertia from prior success can keep them from being company-ending actions, especially if they can be corrected once they are recognized as stupid (or once the people who knew they were stupid from the beginning can gather enough political clout).
Tesla has smart people, just like any tech company, but they're not uniquely smart and they're not immune to the same short-sighted thinking that plagues other tech companies.
Note that I think this particular idea is not necessarily bad for the company, it's just not anything special. It's basically Folding@Home, monetized. If you can convince enough people to leave it on (or force it on and somehow avoid legal/PR issues) then I guess you can save a tiny bit of money on certain kinds of workloads.
No. The very article you link details the difference, and that Counting Sort is often a subroutine in Radix Sort, and that Counting Sort is extremely poorly suited to strings, which is the entire purpose of this sort. And this name is better. So... "no" in basically every way possible.
Characters in a string can be thought of as digits in a base-256 number. Call counting sort recursively on each bucket, looking at the next character in the string. Can you not see the similarity?
There is no one who fails to see the similarity, because they're both kinds of radix sorts. For instance, a counting sort is a degenerate form of radix sort that only does one pass.
But where the classic radix sort is bottom-up, the American flag sort is top-down. That really matters for data where the the most significant portions of the data is likely to be all you need to consider for the sorting. While a top-down sort will have similar performance to bottom-up in the worst case, in the best case it can skip a lot of passes, especially if the inputs are much longer than the longest common prefix.
They're related, but in a very specific way that does not meet the bar for "variation" in my opinion. As you described, counting sort is a very simple procedure that works well only on a very specific input domain. If another algorithm has "more logic" than counting sort itself in order to transform a very different input domain into a format that is suited for making many repeated calls to (a procedure that may be implemented as) counting sort, in a specific pattern, and appropriately coalescing the results of those calls, I think it is appropriate to let it have its own name. Would you prefer to refer to every possible utilization of counting as "that version of counting sort where...."?
American flag sort is essentially an in-place version of counting sort.
That's talking about the single-pass American flag sort. If you use American flag sort to sort strings, you do multiple passes: on the first pass, you sort all strings by their first character, then for each starting character, you sort the strings with that starting character (which are now in a contiguous subarray) by their second character, and so on.
@chowells calls it a “top-down radix sort” below, which is a great description. It also explains the strengths and weaknesses of the two algorithms: radix sort works great for small strings of fixed length (e.g. IPv4 addresses, which can be thought of as 4-byte sequences) while American flag sort works great for variable-length strings like actual textual strings, especially if they don't share common prefixes (e.g. dictionary words, usernames, etc.)
@hinkley pointed out that the recursive version is just a bucket sort, but in-place. Which is also true!
tl;dr:
Single-pass American flag sort = in-place counting sort
Recursive American flag sort = in-place bucket sort
As someone who also bounced off it initially - I would recommend getting through the tutorial area and flying out somewhere in the ship before you put it down for good. Once I started going out there and visiting places, it really grabbed me. The more stuff you scan and read, the more intrigued it made me, and eventually I couldn't stop until I'd unraveled every story thread and mystery the game had to offer.
My very-rough feeling about it from playing around with Stable Diffusion is that it takes about 4x as long if it runs out of GPU memory and needs to shuttle data back and forth from system memory. There are a lot of variables though - on my 3070 with 8GB of RAM, I can get very impressive 512x512 images in about 10 seconds with somewhat low sample counts, or I can set it to a higher resolution and sample count with 2x upscaling and get a really sharp image in around 2 minutes.
Sounds like they could implement local network IoT and reduce their server costs to $0 - but then they wouldn't be able to get a continuing revenue stream.
My anecdotal experience (partially informed by working on the Windows Wi-Fi team) is that iOS/macOS are more aggressive about switching APs. It's a tradeoff - on one hand, you can have disruptive scans like this, but on the other, if it results in switching to a better network during a long period when the user is stationary, it can result in a better experience.
It also depends a lot on what your hardware is, whether you're doing a full scan or a partial scan, whether you have more than one NIC etc, etc.
Video calls aren't hyper latency sensitive, there is already a lot of latency in encoding and processing effects. I'd imagine this could be most disruptive to video games but this isn't a market Apple has done much to work with.
