An Intel Core (i3 or higher) or AMD Ryzen 7000 CPU

Hmmm, no ARM CPU's. Meanwhile the Apple M1/M2/... is blitzing Intel in the compute per watt race.

Only works if a secondary app can get access to a hardware buss or the screen buffer.

Allow me to introduce you to Australia's "Assistance and Access Bill (2018)". The act defines "Assistance" as "we can order any software company to provide us with software we specify (for which they will be recompensed)". Not stated is explicitly is that this covers "the software shall not be reported as a virus or malware, and could be modified versions of (say) your screen and touch/keyboard drivers that report everything the user sees and types". "Access" is code for "thou shall install the spyware we ask for you to create for us via your auto update mechanism, on the devices (and only the devices) we tell you to target". Both a covered by a gag order: the company can never reveal being asked to do either.

As you say, Australia's bill very doesn't break encryption. It also very explicitly bans the government from introducing a "systemic weakness" which could be thought of as something causes all devices to leak data. Only devices listed on a court order can be effected. So it can not be used for sort of blanket surveillance the UK was demanding.

If it wasn't for the bill allowing them to effectively hide nearly all of this activity, it wouldn't be a bad compromise I guess. As it is insists there is almost no sunlight being shone the activities of the governments law enforcement agencies, which in my experience almost guarantees someone is going to find a way to abuse it, eventually.

On social media platforms, algorithms are mainly designed to amplify information that sustains engagement,

The algorithms are designed to amplify revenue. Yes, that might be engagement, but not always. To pick one example, Spotify directs you to songs that maximise the odds of you subscribing. Apple does the reverse - they direct you to songs that cost them the least.

Along the same lines, you could well imagine Google search prefers sites that earn the most clicks on Google ads. I'd be amazed if Facebook doesn't do a similar thing. For example, they could display a few social posts in the news feed that highlight the desirability of some item, followed by an ad for the item itself.

The point being, you're up against an AI's developed by the world's leading computer companies. Those AI's can be far more subtle and devious that you might expect if you thing they are just optimising for engagement.

Microsoft Struggles to Gain on Google Despite Its Head Start in AI Search

Someone is seriously miss-informed. I'm not sure Microsoft has done much AI research, certainly not a lot of any consequence. They purchased a stake in ChatGPT and gave them heaps of CPU cycles, and quid pro quo are now using it in Bing.

But ChatGPT hasn't done much of what I'd call basic research either. They took the the existing design of an LLM (attention heads and whatever) and sunk their time and effort into training it. Don't get me wrong - I'm sure there is a lot of smart engineering went into how they train it, where they get the data, how the curate it and keeping the costs down while they did it. But they didn't do the basic research that came up with the design of the LLM's we use now.

Google did that. They invented the entire shebang in order to power Google translate. Turns out when you translate a word like "plane" to another language you have to know whether it's a miss spelling of plain, or referring to an aircraft, a word working tool, a mathematical two dimensional object a boat skimming the water, or another meaning. To do that you need to understand context it's used in - and thus Google invented the LLM. Not recently either, it was years ago. Then they created customer hardware to power it - the TPU.

I would not be surprised if Google wasn't a decade ahead of Microsoft in AI research.

What did Dell do to piss off the Australian regulators?

Exactly what the article said: mislead consumers. It's illegal in Australia, and the law is enforced.

Retailers of all stripes have been pulling similar shenanigans since approximately forever. The article even mentions that Amazon Prime Day and typical Black Friday sales pull the same stunts.

Correction: retailers in us US have been pulling similar shenanigans forever. But that isn't true in Australia.

My guess as to why is it's because the US government subservient to the corporates. In fact from afar the US style government seems to be more corporate cronyism (ie, a place were money has the votes) than a democracy (were citizens have the votes).

Australia on the other hand is closer to a true democracy. One example: in Australia, insulin is 1/3 of the price of the US, yet insulin is made in the US. A group of companies can't cooperate to change 3 times what it's worth unless the oligopoly can buy a loyal politician. Australia on the other hand plays one off against the other. It's the law, and it's enforced.

Believe it or not, Ukrainians also rely on drones made in Australia. Which must prove just about anybody can make a low cost drone, because with a minimum wages of AUD$23/hr the one thing Australia ain't is a low cost manufacturer. Yet, these are low cost drones.

One reason they are low cost is they are made out of cardboard, and shipped as a flatpak. It's almost as it if they aren't expected to survive their maiden flight. 75 mile range though, compared to 4.5 mile for DIJ.

Somehow, I don't think the Ukrainians will be that concerned by China banning drones.

How can you verify a voter is eligible and casting only one vote, while absolutely not having any way to know what that vote is?

As a point of order, it's possible. In fact there are a number of schemes that do it. It even has a name: end to end verifiable voting. They schemes resemble cryptographic protocols, and like many such protocols what they do seems like magic. But in the case of E2E voting it's understandable magic. I expect anybody with a basic STEM education could read and understand Prêt à Voter which seems to be the most popular.

We don't use them because they are complex and need votes to be record electronically. In advances democracies we have simpler, cheaper schemes that work just as well.

It probably doesn't matter what laws the EU passes when it comes to open source, because open source ignores the law. The prevailing attitude to patents: don't read them. When the USA tried to ban the export of some types of encryption the response was: print the code on a t-shirt and wear it while cross international borders. The UK's proposed encrypt laws: "oh, the UK has some dumb law - why didn't someone tell me?".

An so it will be with this. At worst it might scare some open source programmers in the EU. But only the older ones. The new DVD John's will charge on heedless like all youth does.

The companies and corporates in the EU can't be so reckless of course. If this is a bad law that slows down the pace of technical innovation - well it sucks to be them I guess.

How to provide a simple but satisfying explanation for the depression?

The water is responding to the forces it can "feel" locally. There is no way for the water to feel the absolute height from the earth's centre, so it isn't related to that.

The two forces water can feel are the earth spinning (which leads to the sea level being higher from the centre of the earth at the equator), and gravity. Those two forces combine to create a net force on a body we call weight. At sea level this weight is the same. I guess it's where a mass of 1kg == 1kg of weight. So sea level is really "geocentric equipotential ellipsoid" (to give it it's formal name) around the planet where this force (gravity - spin) has some defined value.

Calculating this ellipsoid (ie, the height above the centre of the earth where 1kg mass == 1kg weight) is a complex process as it depends on mountains, rock composition and lots of other stuff. For example, at the equator it's 21km further from the centre of the earth than at the poles, in this depression it's 100m lower than most places at the same latitude, and at the Himalayas it's higher than the surrounding plains because their mass increases gravity.

We have maps of this geocentric equipotential ellipsoid whose names you may recognise. WGS84 for example is used by GPS as it's definition of sea level. When GPS says "your height is ....", it's not giving you a distance from the centre of the earth, it's your distance from the WGS84 ellipsoid. In this depression it will be 0 if you are standing on the surface of the ocean, despite you being 100m closer to the centre of the earth than somewhere a few hundred km away.

Or just read the part where I said it requires a more expensive manufacturing process vs just using glue.

I'm not so sure about more expensive, as it's hard to reconcile against the fact the lots of manufacturers make phones with replaceable batteries now. Here is one list. The problem for your "it's expensive" argument is they are all cheap phones. (Not being able to buy a high end phone with a user replaceable battery has been a long standing frustration of mine.)

Those expensive phones have one thing in common: the all look expensive, and thus are thin and sleek (no visible joins in the case). Perhaps that's hard to achieve without glue.

As this la forcing you to buy an uglier (but probably not more expensive) phone: yes, that is what it does. But Europe has been forcing themselves to do environmentally friendly things for decades now. Things like recycling packaging. This is more of the same. And to be fair, it works. Europe recycles more than the rest of us.

Besides, student loan forgiveness is a terrible way to solve the problem of sky high tuition rates.

I'm not sure that forgiveness is so terrible, but the student loan idea was a complete cluster. As was the same thing here in Australia. We copied you, amazingly after the problems were becoming obvious. It ended the same way. Predatory firms, offering free iPad's and other trinkets up front to people who could never pass courses in order to get their noses into the government money trough because lump sums were paid up-front. It was a total disaster all round. It was not just the sky high personal debt it created, it was the appalling graduation rates and destruction of educational institutions that had a history of turning out capable graduates..

Make no mistake - this was just anther case of businesses figuring out how to bend the rules to gorge themselves on government money. All that money came from taxpayers pockets, and despite them being called "loans" a lot won't be paid back. If it looks like a duck and quacks like a duck, it probably is a duck. In this case government money was poured into businesses pockets, and they got very little education in return.

They killed the loan policy because it was a disaster, but that does not fix the ongoing pain their mistake is causing. Who do you suggest fixes it? Practically, it can't be the people who were hoodwinked by unscrupulous business operators. The obvious solution is the people who made the mistake pay for fixing it - and that would be the government. And they do that by forgiving the loans.

The best thing Microsoft did for Skype was buy it. Skype was first, but there are lots of very capable replacements now so M$ turning it into crap is not such a big loss. If you want "rock solid and just work (in the west anyway)" go with Facebook or Google. If you want "secure", go with Signal. I have no idea why anybody would go with Microsoft skype or teams (or outlook) - the number of bugs, outages, and UI quirks in Microsoft products just floors me.

In buying Skype, Microsoft freed a couple of very capable software engineers. They found another itch, which happened to be how badly they were getting ripped off by banks when doing international money transfers (one was working in the UK at the time, the other somewhere else). They solved that problem by creating TransferWise. Now called just Wise, they've saved me a heap of money over the years.

Wise has branched into all sorts of Fintech niche's now - like cards, and bank accounts. Mostly free. They have me seriously wondering if I should kill all my current bank accounts and just use Wise.

Really. A terse explanation would work much better.

Maybe we are misunderstanding it. Maybe it's not a coding question. Maybe it's a "can you raise this verbose, meandering but complete piece of documentation typical of 80% of stuff churned by by corporates for their API's" question.

To be fair, it's a must have skill for a professional programer.

Once you have put stress on it, it becomes unreliable.

I do hope the engineers in Boeing and Airbus allowed for that. The Boeing 787 and the Airbus's A350 XWB both are largely carbon fibre. Both undergo repeated compressions and decompressions (presumably 10's of thousands over their lifetime), and both go through large temperature changes. In fact, if you look a videos of the 787 being constructed, it looks superficially similar to the video of Oceangate submarine being made. It's likely where he got the idea from. He, probably correctly, thought there is carbon fibre would work fine in his application if he handled it in the right way.

He could not hire "subject matter engineers" to ensure it was done in the right way, mostly because I doubt there are any. Instead he had have to take the long and expensive route: build prototypes, test them to destruction, over and over again to create your own subject matter experts. Like SpaceX did. (They were one test flight away from going broke before the got something that worked). But you only invest all that time and money if you plan to save it later by building a lot of them. This was a one-off build. It wasn't carbon fibre that killed it. It was the wrong engineering approach for the task.

I don't know what failed - I don't think anybody does. Maybe it was the carbon fibre, but given the lack of testing I'm sure there was a long list of candidates. Even if the hadn't used carbon fibre, given the lack of testing something else would have likely taken them out eventually. Maybe an electrical fire, maybe a battery failure, maybe air supply fouling. There are a lot of things that can go wrong, and you can bet they all will happen at some stage. And when you are 4km down, most will kill you.

I'd like there to be an open source standard for passkey syncing (or less politically correct, cloning)

This will probably come to pass, sadly. The "sadly" isn't about open source. The sadly is you, the user, can't be trusted to manage a passkey, so opening up the standard so random Joe Citizen can see them and copy them is not a step forward.

A passkey is at it's heart just secret private, managed by a few clever protocols. A private key at it's heart is just a secret string, which when it's all said and done a password is also just a secret string. It has turned out humans are atrocious at managing passwords. If we ask them to manage passkeys the won't be much better. We will see them put into speadsheets, copies given to spouses / kids in the same manner PIN's are handed out now, and scammers conning them out of people. While the current Apple / Google implementation stinks because they control your passkey, not you, it's safer because because you can't fuck it up either. Yes, Apple in particular has had their leaks so they aren't perfect either, but they still do a far better job of protecting a private key than grandma can ever do.

Because Joe random can't be trusted to clone a private key responsibly, if I were bank, or a titles office, or PayPal, or lots of other places that must secure things of real value, I would not be trusting any device that doesn't have at attestation pedigree that says "this passkey is managed responsibly". There is no reason why open source hardware could not promise that of course, and in fact I'd say "open all the way down to the silicon" is the most secure promise you could get. But in order to do that, that open source hardware has to be protective of they key as Apple or Google. In particular, it must never give a copy of the key to you, as you can't be trusted. That means is not be as open as perhaps you want. You can know the design, but you can't change the hardware or firmware because if you could, you could get a copy of the key. It can only hand out copies of to other devices like itself, which it knows will be just as protective of them as it is. But to know that, attestation about it's open source pedigree becomes a key foundation stone.

In principle, that's all doable. In practice, we need certification standards that the banks and others are happy with. We need certifying bodies who inspect the open source devices, and sign the attestation key cert to say "we put our balls in the line and say this device respects the standard", and probably 20 other things I haven't thought about. Ideally we will have a "lets encrypt" for open source passkeys. But in reality - none of that exists. Yet.

So for now we are stuck with corporations managing our passkeys, and taking advantage of the situation to lock us into their eco-systems and sell our eyeballs. I can't say I'm happy about it. But for the average Joe, it is better than passwords, so it's very likely what we will end up with in the short term. It's going to take a decade at least for the open hardware designs, standards, certifying bodies, to rise up out of the swamp. Then, and I hope only then, will the open source solution be accepted everywhere.

Until then play with your open source software implementations as they will be the foundation of what is to come, but don't expect it to become your one form of ID that is accepted everywhere.