Because they have a pretty comprehensive range of products, they mostly meet their promises, and they have decent customer service when things do go wrong. When you look at some of the retailer hacks, cart skimming, etc. they are a fairly low risk option for purchasing.

All you have to do as a buyer is pay attention to who is listed as the seller - as long as it is Amazon, or is listed as Fulfilled by Amazon, you should be fine. If you are a Prime member, you can pretty much just restrict what you are looking for to things that are covered by Amazon Prime.

If the seller isn't Amazon or using Amazon fulfilment, then you are taking a bit more of a risk about what they will deliver and the customer service. That said, at least you aren't taking the chance of putting your personal and credit card details into some random internet site.

I don't make judgements about them as an employer, as there is more than one side to it. You might make a case that they can afford to do better, but ultimately, they employ people within the limits of the law - whatever they are doing legally, others can or will do given half an opportunity. If we think the practices are that bad, then we should be campaigning for changes to the law, not just the employer.

It's a decent, reliable default shopping option. I view it that other (online) retailers need to provide a good enough reason to use them rather Amazon. If you don't have a critical analysis of the other options, you may just be rewarding the same or worse behaviour that you wish to punish Amazon for.

In a nutshell: Make a product that fits the users, don't make the user fit the product.

Most people use leather wallets, and wear denim clothes. Leather wallets being especially true as you move up the income ladder (a large part of Apple's market).

Making a credit card that can be damaged / discoloured by a wallet credit card is the antithesis of good design, yet people will swear that it is great design just because it came from Apple.

As you say - for better or worse - AMD won the 64-bit instruction race. (It was certainly the better solution at the time, whether the long-term might have been well served by something more radical could be debatable).

Thing is, graphics plays an important part - whether it is integrated, or the support of external processors. AMD only bought ATI in 2006, and it took a while to really get their ducks in a row, with improved driver support, better integrated chips, etc. By the time AMD would have been in a position to capitalise, they took a misstep with Bulldozer, whilst Intel made great strides with Core.

It's good to see AMD making ground - Zen has moved them back to where they needed to be, and it's the first time I've owned an AMD chipset since the early Athlon days.

Firstly, if people deserve a $125 payment for the damage caused, then that is what they should get if they claim it - not up to that, based on some cap for payments.

Secondly, if there is a cap on payments, you shouldn't be "selling" it on the basis of people getting a high-watermark value. Divide the amount by potential claimants, and tell them they will get at least that amount. Then up the payments if not everyone claims.

And I think we obsess over air travel too much. Yes, it is polluting. Yes, we should minimise the amount of damage it does. But fuel is such a huge part of the cost of providing flights, there is already massive incentive on the industry to reduce the fuel / carbon cost per passenger.

Flying provides for travel that is not easily / conveniently replaced by other forms of transport, when 98% of carbon emissions come from outside of flying and most of that could be far easily replaced / removed.

Flying short distances may seem frivolous - but if it means lots of people buy EVs when 99% of their journeys are short and take an occasional flight versus them buying fossil fuel cars, what is better for the environment?

Yes, we need more efficient long distance travel, and maybe taxes to keep costs high will play a part, but the incentives to move in the right direction are broadly speaking there already.

There is a lot more to gain outside of aviation, that we're actually withdrawing incentives from - solar panel grants, EV grants, etc. - we'll achieve more by cutting emissions from areas that can be very quickly transformed instead of congratulating ourselves on taking only one holiday a year rather than two.

To be honest, it really would depend on how they handle it - right now, we get previews / starts of shows whilst browsing (personally, I find that infuriating - I don't want to be bombarded with the actual content while I'm just flicking through deciding what to watch).

If they replaced that with adverts, but then when you select something to watch it immediately went into whatever it was you selected, I might be OK with that.

But if they put adverts in the way of content - ahead of the show, and even worse, mid-roll - then I would be on my way out

Plus, most businesses don't have their own versions of these tools. In 99% of cases, it makes sense for businesses to use the services of cloud providers.

Microsoft has it's own online office suite, it's own team communication software, etc. - not only does it not make sense for their confidential information to be stored unnecessarily on servers outside of their control, it is simply good for them to eat their own dog food, using the experiences to make their products better.

It's not a driving sim - it's an arcade game. Within the parameters of what it is, it's... OK. A bit overpriced, but a mildly fun diversion.

No they aren't - in fact both Google has fixed the browser detection to work with Microsoft's error, and Microsoft have pushed a fix to the error in Edge.

It looks like simply optimization - sniffing for "Edg" is more reliable than sniffing for "Ed", and faster than looking for "Edge" (about 25% faster, which is negligible on single access, but significant when you are handling millions of requests.

So it has classified the browser as something it needs to check the version of - but has been unable to parse the version (because they weren't expecting the mistake). No version number = old UI.

As Edh won't get the agent classified as Edge, they will look for something else - the Mozilla, or WebKit or Safari, or Chrome statements in the user agent string, and then treat the browser as one of those.

Or they want to support devices / apps where delivering modern interface would cause a problem, so they default to the old interface and sniff for anything that is broadly compatible for the modern experience.

There is Opera, and Vivaldi and about a dozen other Chromium based browsers out there. How many of those are blocked from getting the modern experience on YouTube?

As I explained, this is most likely NOT about blocking Chromium browsers, but that they have identified it as Edge - which, until very recently, will only have meant an EdgeHTML based renderer. And because the algorithm couldn't parse what version it is, it wouldn't determine that it was a minimum version required for the modern experience.

Anyway, Microsoft have updated the Edge Dev channel, and the user agent is now correctly reporting Edge/76 rather than the spelling mistake. And guess what - Google has already deployed a fix to YouTube so even if you manually set the user agent back to the typo Edg/, it is delivering the modern experience.

It was simply a spelling mistake by Microsoft, and a quirk (an edge case!) in Google's detection algorithm, that both have fixed - no grand conspiracy. Nothing to see here.

Presumably this is a quirk of an optimization / code path.

Look at it this way - first thing any parsing is going to do is identify what type of browser it is, then figure out what version it is.

It's common for user agents to contain multiple identifiers - in the case of Edge, it contains the strings for various other browsers (Mozilla, Chrome, Safari, etc.), so you need to look for something distinctive.

Looking for "E", and possibly "Ed", is not sufficient to identify the user agent as Edge. It's likely the case that searching for "Edg" is the fastest way of reliably identifying it as an Edge user agent. If the code isn't identifying Edge, then it will identify Mozilla, Chrome, or Safari from the same user agent.

So then you try parsing for the version number - which you know always comes immediately after "Edge/", except that parsing isn't going to work when the user agent doesn't include "Edg/".

This would leave the code identifying Edge, but without a recognised version number - which wouldn't then pass a minimum version check.

And that's how it is entirely feasible that an optimized browser detection can fail to identify only one type of spelling mistake as a compatible browser version, without it having been a deliberate action to screw up one browser.

YouTube is an important platform for Google, that they available on many browsers and devices. So what seems more likely - that it is a quirk of how an error / change in the browser agent interacts with an algorithm, or that Google has gone out of it's way to hurt a browser preview that will only have a handful of users, while Firefox, Safari, Opera, Vivaldi, etc. all work correctly?

Almost - Microsoft managed to change the identifier from "Edge/" to "Edg/", and Google's algorithm can detect what version of Edge it is because they are looking for "Edge/".

Note that you don't even need to drop the Edge version number down to a recognised number - you just need to correct the typo in "Edg/" to get the modern experience on Edge Dev.

Clearly, the modern experience is only enabled when a compatible user agent is found, with the old interface the default (for maximum compatibility) - one small typo is enough to prevent it being a recognise user agent.

In the current Edge Chromium, the user agent is listed as:

Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/76.0.3800.0 Safari/537.36 Edg/76.0.167.1

Note the bit at the end "Edg/76"

The "old" edge has a correctly spelled "Edge/XX" user agent.

If you enter a custom user agent and simply correct the spelling to "Edge/76", then you get the modern YouTube interface.

Clearly the user agent detection algorithm for YouTube recognises that it is not Chrome, Safari or Firefox (as the user agent might hint at), but it also can't recognise and parse the Edge version, and therefore it doesn't see it as a browser that is capable of the modern experience.

Yep.

I can see that it might be potentially useful for a componentized website in a way that iframes aren't.

But browsers already have way too much useful functionality that doesn't have sufficient limits to prevent it from being horribly abused to the detriment of users, where ultimately - regardless of how useful it can be - we would be better off it didn't exist and therefore couldn't be abused.