m4nu3l

Also, not all language models are transformers. You can have language models based on diffusion models or state-space models, or any other model that can be used to model sequences (so all of them, as sequences are just functions). Transformers are just the ones that are most common and successful today.

The study seems to be strictly for TV watching. I mostly watch science/math/history/economics related videos on YT. I prefer the more in-depth ones; I also watch gaming/movies related videos sometimes. I don't think it's nearly the same as watching TV. But I don't watch hours a day of those either.

When I find articles like these, I remember I haven't watched TV for almost 12 years, except when at someone else's home.

By this, I mean I never turned on the TV to see what's on the television. I used the TV as a monitor to play games or watch movies from streaming services at times.

Given that I moved to the UK after I stopped watching TV, I barely know which channels even exist on the TV.

I do watch a lot of stuff on YouTube. But it's things I chose to watch because I find them interesting, and I don't find most of the stuff on TV interesting enough.

And the interesting parts, usually, you can find them on streaming services and only watch those whenever you want to watch them. Even scrolling brain rot on Instagram seems less of a waste of time. At least the algorithm knows the type of brain rot I like.

I wonder how many people still watch TV today.

> When getting something out on the market, they woudn't take this risk and still have a commonly used port on the side. The consumers as well would prefer to not take the risk of buing a device forcing them to use a port no other device is using.

Like, people didn't buy Apple products before they were forced to use USB-C? So why did Apple not want to use USB-C if that caused them to lose money and sell fewer phones? Also, if this were true, once most devices support USB-C (like they already did before the regulation), the standard would be self-enforcing and require no additional regulation.

> I apologize, but I only skimmed the rest of your comment, because I believe it rests on a flawed premise in the first line.

EDIT: Your statements seem to be self-contradictory, so it's not even about facts. "We need a regulation because companies make and sell products with different ports" is in contradiction with "Companies wouldn't take the risk of not having a common port, and consumers won't take the risk to buy them".

Also, mine wasn't a premise. My claims are not statements that follow one from the previous, but a conjunction of negated statements laid out in chronological order. So all of them must be false or not relevant for you to be right.

The point is, if the regulation works as intended, it consolidates USB-C as the standard and slows or blocks new types of ports from being designed and tested. If it doesn't, then it's useless. You can't have it both ways.

Imagine a company that wants to design the next type of port. Normally, it would just do R&D and engineer the product, then put it on the market with that type of port. If the device only has that type of port, people must use it and provide feedback by choosing to keep buying the product (maybe they'll also make it explicit that they like the new port). If this phase is successful and the company opens the specs and other companies adopt it, it might become a de facto standard.

Now consider doing it with the current regulation. You need to do R&D, but the design will require higher costs because you have to make the device work with the old USB-C, too.

Maybe it's not a high cost for the company, so they do it anyway. People might use the port much less because some may not even try the new port and stick to the USB-C for simplicity, or because they are accustomed to it.

So the company won't have the same strong feedback. But let's assume they manage to make most users happy to use the new one because it's so much better. Now they open the standard. Other companies have to pay the additional engineering costs to add the other port. Some might be able to afford it just fine, while others might not want to add the complexity and costs of two ports.

Let's say, however, it spreads anyway. Now you hope the EU will accept it as a new standard so you can ditch the now-old USB-C standard. But the EU might just say "NO". It's very hard to predict if the government would be happy with it.

Given the compounding of these effects, the result is most likely that the company won't even think about wanting to develop the next standard.

The probability of a new standard happening is much reduced. And as I said before, if it weren't reduced, then the regulation would be useless to begin with, as people would use and test different connectors regardless of the USB-C support.

As I said in the root comment, I don't care too much about the connectors not evolving, but it just goes to show that people don't realise that there is no free lunch. Add constraints, and you add costs and slow down progress.

You can't force companies to add value with no costs, and value is also subjective, so the government often doesn't know what is valuable to people. At most, the government might know what some people say is valuable to them. But this is not the same thing.

Very few people would have said smartphones were the future, and they would buy one in 2007. Democracy is a very bad way for economic and technological progress to happen - unless it's for some military technology for which the government is the only customer.

What's interesting to me is that this is a bit like how the USSR worked.

They were relying on the West for high-tech tractors and even to build factories (I know Ford helped them build some factories for a part of the factory output). I also know they used to look at the allocation of goods in the US as a starting point to decide what to produce, as they were basically running blind to what people needed or wanted.

> you just have to provide USB-C option as a fallback.

Given you must have a USB-C port, the design would still be limited by the USB-C form factor and adding another port would even further increase the complexity, costs and weight of the design. This will still strongly limit the spread of new designs, and I'm pretty sure it's effectively equivalent to mandating just USB-C for smaller devices.

> On the other hand, USB-C wouldn't have become a true standard if no-one forced Apple's hand

I don't want a state-dictated standard like this. What you're saying is that because some people want iPhones and they want them with USB-C, everyone else must forgo the possibility of having a better type of connetor until "we" (Is it the majority? I don't even think the majority uses iPhones in Europe) feel like having a new one (at which point the progress has been delayed anyway and you'll also get the initial problem again). I find the premise quite capricious.

> I don't have this particular problem, so it doesn't exist!

No, what I said is that you could find devices with USB-C in all the categories that are now regulated. This means it was pretty easy to find devices like that if you really valued USB-C. Of course, if you wanted an iPhone but you liked USB-C, you would have had a problem. A problem that is much less worse than blocking progress.

> Which is why the law can be simply amended as soon as such a standard emerges. If the industry figures out something better than USB-C, pressure will build on the council to do so. This is nothing but a straw man.

You totally ignored what I wrote, or you didn't understand it. No standard can emerge if you can't test it on the market. You can have a bureaucrat choose the next one from some proposal. It's not the same.

This is exactly what I'm talking about. A short-term view of the world, progress and technology.

All my devices supported USB-C before the EU regulation. But if I wanted to buy a device with a new type of connector, I should have been able to. This is how the USB-C came to be and how any new standard in hardware happens. New technologies are made and just sold, and if they are proven to be superior to others in the market, they often become standards.

The USB-C standard is not the best standard that can exist from now to the end of the universe, but if this discovery process is blocked, we will be stuck with it forever, which, of course, will also constrain the design and engineering of devices in other ways.

It's the same fundamental flawed thought process that has made the EU reliant on the US for a lot of services.

The education system has failed in the EU, but in a different way than it has in the US.

I realised this when people thought mandating the USB-C connections was a good idea because "it is the best standard". I didn't think the mandated connector was a huge deal per se, but it made it clear to me that there is a flawed thought process behind EU regulations. And this is a big deal.

Many things are not really understood in the EU. The majority don't seem to understand free speech. The EU has an article about free speech that clearly states there is no free speech, but people point to it when they claim there is.

Not sure what you are referring to. Can you elaborate?

Just to be clear, my definition of free-market is just that there is no centralised authority that can use force to set prices/quantities/quality/type of services offered. Of course, the fact that the employer has to offer health insurance in some cases is part of it not being a free market. But there are more fundamental things that make the US healthcare very far from being a free market. The first one is that the supply of doctors is capped in quantity, not just in quality.