Peer to peer viewing can only go so far. Some people, when they put a video out, get hundred of thousands of view in the span of a few minutes. This works relatively well on youtube, with a very large CDN (and probably some heuristics for big accounts). It is enough to hinder “smaller” platforms like dailymotion. It would just be a terrible experience on peertube as it is now, unless the creator preemptively mirrored it in many, many places beforehand.

Accessibility, usability, scalability at very, very large scale, actual searchability, and actual return on investment, because some people actually get money from youtube?

Actually, peertube, depending on the instance and the popularity of the content, can be incredibly frustrating for a viewer. And it can be frustrating to the content creator. Some people are quick to dismiss minor (and less minor) annoyances, are able to look for fixes, and so on, but for almost everyone? The experience is nightmareish, with incertain returns (or no returns at all, as it stands).

Once you fix all that, you might have a chance to convince larger entities to move to peertube. Well, more realistically, to host their own instance. Well, more realistically, to host multiple instances, because really some people would hammer the platform down with each video. See the issue yet?

Lives gets destroyed, everyone pay for it, money is diverted from actually useful stuff, but all this is a low price to pay for a handful of mafia bosses to get even richer.

Not only in Paris. Saw some of them in Marseille too. At first I thought it was some kind of “anti-ad” joke, where someone was giving them a bad rep because of how stupid it looked.

Guess it was not a joke.

So-called “social” networks can have three main issues: technical (they have to work), leadership (they have to not be dickhead), and users (they also have to not be dickheads).

The first point can be handled with competent people, consensus, open source contributors, etc. (assuming no dictatorial management).

The second point can probably be handled by having a handful of decent people, transparency, accountability.

The third point, which is basically the thing that makes the content on the service… is still people. If people were obnoxious on twitter, they’ll be obnoxious on bluesky, mastodon, and whatever else shows up. It’s almost inevitable.

It’s also why decent moderations tools are needed, which brings the question of how to do decent moderations tools that are not too extremet but still remains useful. This is not an easy task (and to my knowledge, there’s no general solution to that).

Bots showing up is just the icing on the top. Without a pretty aggressive vetting system for accounts, there’s not much that can be done from the service itself.

Given the general ambiance, I guess smaller community and services tailored for them might come back, the way we had tons of different forums back in the days. It might be a good solution; some form of SSO across many services to make people reachable, but no general, shared stream of messages as we have now.

tl;dr: it’s not a technical problem, it’s a people problem. So it won’t be solved by technical solutions.

I’m worried about anything that can be controlled through a third-party online service. The amazon doorbell thing is a prime example of what can go wrong, but it can be more subtle, too. And I’m not even talking about obsolescence. Frankly, I’d still be worried if it was a self-hosted, properly configured system where I’m the only one with a legit access.

I understand the convenience of all this. I also have to deal with the risk balance of security vs convenience, which causes me to not tolerate that “too tired to go across the room” justifies “a third party have full control over my doors, lights, heating, ovens, etc.” (not shooting fire at you, see this as a generic example).

The bare minimum would be a fully self-hosted solution, which is possible, although difficult because hardware manufacturers don’t always play nice. And even then, proper, secure setup and maintenance is not for everyone. In the meantime, yeah, I’ll have to move myself when I want to turn on my dishwasher.

Though I’ll admit, I have some lights that are controlled wirelessly… my old phone have an IR port, and they have IR remotes… Technically, an attacker could probably turn them on/off/change colors from behind a window :D

I’m warry of electronic, wireless, and sometimes third-party cloud dependent services, having a say in how I lock my doors or control heating.

I’m a bit old fashioned, but also have to work with solutions where considering the consequences of a compromised entry point is vital. I’d be ok with a way to check that the door is locked, but something that can lock (and, so, unlock) my door remotely? Not a chance. At least, not for a place a value.

Probably a security update to try and keep it from being part of a botnet maybe?

Then we’re back to the same question. At what point a grill have anything that could be part of a botnet :D

Grill, Dehumidifier, Air con, Fridge, Dishwasher, Washing Machine, Lightbulbs, Ovens, Doorknob…

None of that should be smarter than “press button, get action”.

And where do you dump the paper towel then? The idea behind a jar is that there are facility that takes that grease and (supposedly) won’t dump it in a trash field, which is pretty much bad too.

A crash? I’d say we’re in a plane from which the pilot decided to press the “eject wings” button.

The whole industry is projecting something like negative $200B for next years. They know it’s not worth the price.

People don’t care. as long as they can get their infinite scroll with funny picture, they’re happy.

But you can. Hopefully, you know how your language is called in your language, right?

I’m pretty sure nobody’s doing that based on geoip. Client-side, the browser exposes the user’s languages choices. And server side, the HTTP header can help. But geoip is totally unreliable, even a broken salesman would not sell that as a feature.

Well ok they would sell it but get a very heavy glance from the dev team.

The label for the language picker is an issue, but the choices themselves? In the target language. You want French? You pick “Français”. You want Japanese? You pick “日本語”. You want english? You pick “English”.

Supposedly, if you’d rather have a website in a given language, you must have some level of understanding of that language, and picking its name should not be a challenge in any case. If you somehow change a site/app to a language you don’t know, as long as you can identify the language picker, you’ll be able to change to something you understand.

It does leave out the case of a user wanting to change to a language they do not understand, but I do not care for those.

Flash was a security nightmare all round, not counting the security flaws. It was just designed without any security features. It was also terribly inefficient at its core job, that was supposedly vector animation. It filled a gap in a time where browser and standards where not that advanced.

Over time, Flash issues where never resolved, but the bloatness of the software kept increasing. Along the way, HTML got better specs, JavaScript got vast improvement, especially in everyone adhering to roughly the same standard (thanks microsoft for finally caving in…), and so the flash interpreter was highly redundant with the browser itself.

For a while flash editors could export in HTML5 and you’d get roughly the same result, but with a fraction of the resources requirements, so naturally there was little incentive to keep the flash player around.

I’m not sure if “killing flash” could be attributed to their author, or to the loss of interest.

Also note that alternative flash players exists to still play older swf files, and some sites uses them alongside with plain video conversion for flash animations that weren’t dynamic.

Yes. But if you live in the future, you have to wait for dozens of dozens of intermediate to do so! Great!

Likely, you’ll see the first frame only on older software. Encoding animation in a dedicated animation chunk and using the base spec for the first keyframe sounds like the sane thing to do, so they likely did that.

I’m not going to look into it now, because I would then have to implement it. :D

The PNG format is made of chunks that have determined roles, and provides provisions for newer “standardized” chunks alongside the custom chunks it had supported until now. It is likely that PNG made with newer software that does not use new features, or uses only additional features, will remain readable by older software to some extent.