> I don't know why Americans are so confused by normal roundabouts.

Start with lack of familiarity. Imagine you're a tourist, and you're visiting a place for the first time, that is far from home. You're already unfamiliar with the specific roads and intersections, because you're from out of town, so navigating is already a bit of a challenge, and now suddenly you encounter a whole new *type* of intersection that you've never seen before, and the rules for how it works are completely, totally different from anything you're familiar with. The locals are all zooming through because to them it's normal, so you feel a social pressure to keep up, but you have no idea where you will end up if you do a particular thing, or what you should do in order to end up where you want to be. Your GPS gives you instructions that you don't understand, using terminology you're not familiar with. How are you going to feel about this experience?

Right, that's a separate issue, but it's an important issue nonetheless.

Prioritizing durability over ease of repairs makes a lot of sense: why spend money repairing things, when you can just make them so they don't *need* repairs in the first place? That tracks.

The problem is that they're also prioritizing a lot of other things over ease-of-repairs. Thin-ness and sleekness, lower manufacturing cost, speed, battery life, you name it, almost everything gets prioritized over ease-of-repair. There are market-related reasons for this, and most of them are rooted in the fact that Apple have deliberately positioned themselves as a luxury brand for upper-middle-class consumers. This inevitably means their products are NOT going to be the best option for technical users, or for budget-conscious users, among other demographic categories.

Yes, it's such a tragedy, because the world doesn't have anywhere near enough Tetris implementations, and nobody else has any idea how to make them.

Depends on the glue. If the glue is made by old-fashioned knackery, then no, that's not vegan.

But there are plenty of plant-based glues, and even more synthetic ones. Gorilla glue, despite the name, is a fully synthetic polyurethane adhesive, for example. No gorillas are harmed or otherwise exploited, except possibly for the advertising campaign.

The peck, like the bushel and half-bushel, is an inherently imprecise unit of measurement, because it's a picking-basket size, and picking baskets are *intended* to hold things that come in irregular shapes and sizes, like apples or cucumbers or rocks or whatever. You can pick into the same basket a thousand times and never get precisely the same volume of produce twice. Consequently, if you have two peck baskets and one of them has an interior volume 5% larger than the other, it doesn't matter, at all. One of the baskets isn't wrong or bad; they're both perfectly fine for their intended purpose. That remains true even if you are using the baskets to sell produce by the peck, because the minor difference in basket size is smaller than the amount of imprecision inherent in measuring fresh produce by volume. If you need to be that precise, you sell it by weight.

Also, the correct answer to the original question is, all of them. You should eat all of the rocks, every day. HTH.HAND.

TV and radio commercials kept running whether you were watching them or not; if you switched stations for a bit, the ad you didn't like would finish, rather than continuing to play when you switched back. And yes, all the stations ran commercials at the same time during the between-program slots around the end of each hour, but at least the other channel would have a *different* commercial, and also, those commercial breaks were relatively predictable in duration, so you could use them to do things like go to the bathroom or get a snack. Compare to when a live streamer goes "I need a five-minute break, so I'm going to run a couple of ads now...", which hardly anybody begrudges, not only because we all understand that the content creator may need that break, but also because when content is live, people don't want to step away and miss something, but everyone needs breaks occasionally, not just the content creator. Even live theatre has intermissions. Watching pre-recorded content and having the ability to pause at any time, changes the psychology of ad breaks.

Also, TV and radio commercials didn't cause lots of technical glitches. (MOST of the annoying bugs in YouTube are related in one way or another to the ad-loading system. It's terrible.)

If you recorded a TV program on a VCR and went back and watched it later a few minutes at a time, it didn't insert new advertisements every time you resumed playback. Similarly, rewiding or fast-forwarding through the content did not cause the VCR to load in an additional advertisement and make you sit through it before you could continue finding the place you wanted to find. The quantity of ads in the broadcast was fixed and did not increase if your viewing pattern was different than the broadcaster imagined. You could even fast-forward the ads or, if you saw one that was funny and wanted to show it to somebody else, you could rewind it and watch it again. And advertisers did sometimes make ads that were worth watching again or showing to someone. Occasionally they still do, but YouTube doesn't do nearly enough to encourage that.

> downloaders like yt-dlp would fetch a video that includes ads, and YouTube could *fighting* such tools

Could stop fighting such tools. I had the word "stop" in there but apparently lost it while editing my comment. Sorry for any confusion.

> I don't get why they didn't just insert the ads into the video stream directly on the server side.

That's *one* of my recommendations.

> That would make them harder to skip.

It would make them harder to skip *automatically*. But it would mean that the normal play-position controls (e.g., the 5-second skips forward and backward, etc.) would work the same for the ad as for the rest of the video. This makes the UI *consistent* which is a good thing for the user. This is probably WHY they don't do it, because they're afraid everyone will skip-forward five-seconds at a time through all the ads. And yes, a lot of people probably would. I still think it's what they should do. On the other side, downloaders like yt-dlp would fetch a video that includes ads, and YouTube could *fighting* such tools and possibly even make "download for offline use" an official feature and pop a download button straight into the UI, why not. Which would be really nice for people with spotty connections, among other things. (And yes, technically inclined users would be able to edit the ads out of the downloaded videos, but realistically, only a small percentage of viewers would do that, and only if they were planning to watch the thing more than once or show it during an event, and the latter would be something YouTube and/or the copyright holder of the video could go after with lawyers if the event is sufficiently large and high-profile to matter.)

> it would suck for viewers

Honestly, I think it would be an improvement for users who don't use an ad blocker, for several reasons.

1. As mentioned above, the UI would now be consistent, and that's a big win.

2. In the event you see an ad you actually want to go back and watch again (wholly or in part), e.g., because it's funny, you can easily do that under the proposed system. Under the current system, you can't. This is cool on its own but also makes it significantly easier to incentivize advertisers to make ads that are actually entertaining, something I feel strongly that YouTube ought to be doing anyway.

3. In one fell swoop it kills about 95% of YouTube's current large raft of really annoying bugs and performance problems. It automatically eliminates the bug wherein the ad system tries to load an ad and fails for unknown reasons (yes, this happens frequently even when there is no ad blocker involved) and then video playback stalls indefinitely waiting on the ad that is never coming, and the whole page has to be reloaded. It fixes the bug wherein playback is paused for too long during an ad and another ad tries to load during the ad, causing the UI to lose track of how much of the actual video has been played, and the ad just loops forever until the page is reloaded, which places the player back at the beginning of the video. It eliminates the situation wherein leaving a video paused for more than a few minutes guarantees you new ads when you come back, even if you just watched ads right before you paused. It fixes most of the bad interactions between the YouTube interface and tabbed browsing (in particular, when the user queues up multiple videos in tabs, and then wants to watch them in turn). If the ads in each video aren't changed out too often, it allows users to engage with the ads in the comment section. And finally, this also allows content creators to easily see what ads are being shown with their videos, and opens the possibility for the site to add a mechanism for them to object to certain ones (e.g., I've seen the algorithm pair ads for alcohol with content that was obviously intended for underage viewers; this would allow that sort of thing to be reported and fixed, on a per-video basis).

So I think it would actually improve the situation for users who aren't using an automated ad blocker; and as for users who do use an ad blocker, YouTube doesn't have any interest in preserving their experience, so that wouldn't be a consideration from their perspective.

There is one small downside. If ads are allowed to be however long the advertiser wants, which I think is currently the case, this would break timestamp links. The obvious solution to that, is to only accept ads that are certain lengths, and when swapping them out, always pick ones that are the same length as the ones that were in the video before. Television used to have standardized lengths for commercials, so obviously it can work. The same advertising campaign can even have multiple versions of the ad, for various-length slots.

No problem. There are even more non-dairy options for the glue, than there are for the cheese itself.

I don't know if I've met any humans quite that dumb.

I *have* met humans dumber than the smartest dogs I've known, and attempting to interact with those humans was profoundly disturbing.

Because computer software (such as an LLM) is written by humans and designed to interact with humans, there's a familiar interface that feels computer-ish, and so our expectations are tempered and interacting with it isn't particularly disturbing. But I suspect, if it were using speech synthesis and embedded in a high-quality faux-human body, and given control of gestures as well as speech, it would probably freak everyone out.

The red squiggly underlines in word processing software and the like, are fine. That's not what we're talking about. Users can look at a proper noun with red squgglies under it and go "Oh, haha, La-a's name isn't in the spelling dictionary, gosh, I wonder why not", and there's no need to do anything about that, because you know what's going on and nothing has _happened_ other than alerting you to the non-dictionary-word nature of the name (which you were likely already aware of). That's fine.

What we're talking about here, rather, is smartphones, mostly, and their extreme and pervasive use of automatic corrections that many users would prefer to turn off but can't figure out how. You type one thing, and your phone automatically changes it to something else different, without consulting you, so then somebody gets a text asking "can you pick up some quips?", and it's really not obvious that the thing that's wanted is cotton swabs, because the message they got isn't what the other user typed. There are multiple subreddits entirely devoted to the tragedy, hilarity, and heartbreak resulting from misunderstandings that stem from this issue. The real problem here is that the autocorrect feature is on by default, and it shouldn't be. If users had to turn it on in order to use it, they'd be far more likely to know how to turn it off, so then when they try it out for fifteen minutes and find out that it causes more problems than it solves, they could just switch it back off. Given the frequency with which it causes larger problems than anything it was designed to solve, it really really should be off by default.

Or, people could learn how to pull up Google and type something like "How do I turn off autocorrect on Android?" But we're talking about regular end-users here, so that might be asking a bit much.

A company that can't manage a basic substring search, is going to implement an advanced neural-net-based search. Clearly, this will solve all of the problems.

It's like a building contractor whose garden sheds are terrible, saying "That's fine, we'll just start building skyscrapers instead, it'll be great."

Most schools will have teachers who can make good use of this for some classes, for a whole collection of reasons. English teachers will want to teach college-bound students how to properly cite sources when generative AI is involved. IT and possibly some art classes may want to teach students how to construct prompts that get useful results out of generative AI; many of the details will be obsolete by the time they graduate, but it's still useful to have a foundation to build on. Some teachers may even want to introduce students to the limitations of generative AI, what it can do and what it can't, and possibly even touch on how to spot AI-generated content. So it's useful for schools and teachers to have access to a real-world example of generative AI for these purposes.

The caveat is, if somebody in upper-administration at a school district buys too much into the hype and tries to enact some idiotic initiatives that try to make generative AI do more than it really can, teachers will definitely suffer, and students might as well. But, frankly, generative AI isn't needed for that. If somebody in a position of power is inclined to push idiotic initiatives, they'll find some with or without AI.

Eh, just about everyone I know either parks in a driveway, or on-street parking. Most houses do *have* a garage, but there isn't necessarily enough empty space in it to fit a car. Older homes usually have a single-car garage that was built in the first half of the twentieth century, so it's just _barely_ large enough to qualify as single-car in the first place, and then previous home-owners added a bunch of shelves and racks and cabinets inside for all the things you have to store in there: garden tools, snow shovels, snowblowers, leafblowers, lawnmowers, hedge trimmers, weed whackers, bicycles, sleds, wagons, old buckets of paint, leftover patio bricks, lighter fluid for the grill, camping gear, ... There frequently isn't room in the garage to park a car and open the doors and get out. That's without adding a charging station in there (and last I checked those were still somewhat bulkier than they really ought to be, though I imagine that will get better over time).

Newer homes do often have a two-car or even two-and-a-half-car garage, which allows you to fit one of your cars in there, but the other cars still have to park in the driveway or on the street. I have never personally known *anybody* who had enough garage space for *all* of their cars, unless you count people who don't have a car at all.

Some of the other poster's other points, are less defensible. Though it's still true that the higher cost of EVs puts them out of reach of a lot of folks for now. In the case of the majority who buy used cars, this is a problem that would eventually solve itself, *after* the people who buy new cars all start buying EVs. That, however, is precisely the segment of the market that seems to be hitting a plateau in the last few months. Speculation as to why that is happening and how long it will take for the market to move past it ... well, that could be a whole thread on its own.

Although the other poster argued badly, the point remains that an awful lot of the people who really wanted an EV, seem to have already got one at this point, and the manufacturers are having an awful time selling them. They aren't hoarding them because they want the tech to fail. They aren't refusing to manufacture them. They're making them, they're offering them for sale, in many cases at a loss, and the bloody things are accumulating in the dealers' lots because they're a glut on the market at the moment. This has been going on for months and is currently still getting worse.

Incidentally, the problem that the article actually talks about, related to the spin of the projectile, is a relatively straightforward problem to address. I assume they're going to solve that one within a few months, and have a second go at the demonstration shot, which may very well be successful.

But when that happens, it doesn't necessarily mean they have a useful weapon. They would still have to solve the equipment-wear problem. And there are some other issues as well, but that one's the biggie.