Well. It seems that Google has been cowed by -- or now is under the complete control of -- fascist filth.

Post links to viable substitutes for Google's various services here.

Yes. So it should be legal for them to make that change. It should also inform everyone that YouTube is even less trustworthy than you previously thought. (Unfortunately, it probably won't.)

There are lots of domains were physical evidence is either missing or impossible, yet where many people feel the need to have certainty.

Actually, the space is even larger than that. Every area of expertise implies an area that is not being examined, since people have only finite intelligence and finite time to explore. So...I "believe" in the EWG multi-world interpretation of quantum physics (with a few modifications). This is a belief, because I'm nowhere near expert enough in the field to have detailed knowledge. I *do* acknowledge that there are other interpretations that fit the existing data equally well, but I find them...distasteful.

Also, I believe that my wife was a wonderful woman. This is not based on globally accessible knowledge, partially because "wonderful" is not well-defined.

Etc.

You are wrong. AI has done mathematical proofs that were new. It *can* only be original by combining existing information into new patterns, but if the "rules of inference" are good, this can allow it to create something new and good.

OTOH, you are partially correct, in that it can't derive anything that wasn't already implicitly implied by the existing knowledge.,,because it can't currently run its own experiments.

N.B.: This is a comment about "AI" not about pure "LLM"s. Pure LLMs are a lot less reliable, because they've been designed to never admit that they are uncertain. And because they've been trained on the Internet.

What are the analogous problems in China?

During the early years the laser was called "a solution in search of a problem". Don't try to estimate what current AIs can do by the applications that they are currently shoe-horned into.

OTOH, every speculation as to how AI will develop further is *speculative*. That explicitly includes the speculation that it will not get any better or more efficient. (And I'd call the speculation that "we've reached top AI" at least as silly as "AGI will show up tomorrow and solve all our problems".

Nazism isn't an internet insult. It was an actual, real thing that happened in this world, with an actual, real ideology, and the core of this ideology was racial-based social Darwinism.

BTW, re: the Congo in particular: the most common traditional type of fishing is basket fishing with woven funnels suspended in the rapids. You sure as hell better know how to swim if you want to do that.

Famous angler Jeremy Wade referred to the local Congo fishermen as nearly suicidal, just diving into the rapids to get nets unstuck and the like.

SIGH.

There were 10 people chosen and people with dark skin in the USA make up about 1 out of 8 Americans.

1 in 8 is 12,5%.

African-American without mixed race in 2024 is estimated at 46,3M, or 14,2%
With mixed race, that rises to 51,6M, or 15,8% of the population.
Some hispanics have dark skin, some light. In 2023 there were 62,5%, representing 19% of the population (though there's a small overlap with black - doesn't affect the numbers much).
In 2023, Asians were 25,8M people, or 7,7% of the population. This is again a diverse group with mixed skin tones (for example, the Indian subcontinent)
In 2023, there were 1,6M people (0,49%) of pacific island ancestry and 3,3M native Americans - again, mixed skin tones.
People of Mediterranean European ancestry often have so-called "olive" complexions.

With a strict definition of dark skin, you're probably talking like 1 in 6 or so (~16,7%). With a looser definition, you could be talking upwards of 40% or more of the population.

The chances of the 10 people to be a perfect representation of the racial demographics of the USA is quite small.

Here are the actual odds of selecting no dark-skinned people at different population percentages being "dark skinned", by one's definition of "dark":

15%: 1 in 4
20%: 1 in 8
25%: 1 in 17
30%: 1 in 34
35%: 1 in 73
40%: 1 in 165

Then consider that NASA astronauts are required to pass a swimming test

It is not a test of swimming prowess, just of an ability to not drown. You have to be able to do three lengths of a 25-meter pool without stopping, three lengths of the pool in a flight suit and tennis shoes, and tread water for 10 minutes while wearing a flight suit. This is not some massively imposing task. You don't have to be Michael Phelps to become an astronaut.

and as a general rule those with African ancestry tend to have less stamina in swimming than those with lighter skin

Yes, white athletes tend to have an advantage in swimming. A 1,5% advantage. While a 1,5% advantage may be of good relevance at the highest level of a sport, it's hardly meaningful in a "can you tread water with a flight suit on" test.

Think of the different races as just really big families

That is not how genetics work, and is instead the pseudoscience that drove fascist movements, and in particular, Nazism.

There is far more genetic diversity within a given "race" than between them. Certain genetic traits tend to have strong correlates - for example dark skin and sickle cell anemia - but that's not because races are some sort of genetic isolates, but rather for very practical reasons (dark skin is an adaptation to not die of skin cancer in the tropics, and sickle cell disease is a consequence of a genetic adaptation to not die of malaria which also happens to be found in such climates). But the vast majority of genes don't have such strong correlates.

The concept of "race" as a distinct biological category is not supported by modern genetics.

If we are to ignore skin color and just put one big family up against another big family on swimming ability then just due to random mutations, perhaps some Darwinian selection way back in the family tree, one family will swim better than the other

The main "racial difference" in swimming ability in the US is "inherited", that is, parents who don't know how to swim tend to not teach their kids how to swim. As a result, white children are 56% more likely to receive swimming lessons than black children. One can expect that to directly correspond to an advantage in adulthood. But again, the ability to tread water is not out there knocking 90% of astronaut candidates out of the race - especially given that astronaut candidates tend to be athletic and motivated to learn new skills.

People with light skin tend to have ancestors that had to go fishing for their protein

Utter tripe. Fish consumption has no correlation with skin colour. How much fish do you think your average herder or plains horseman ate? And fish is massively important in much of Africa - in coastal areas (Gabon, Ghana, Sierra Leone in particular note), along the Congo (it's literally the world's largest river, people have been fishing it since time immemorial), Lake Victoria, Lake Chad, the Niger Delta, etc etc. What sort of racist stereotype world are you living in where black people don't fish?

Having a large candidate pool relative to the number chosen makes it easier to show bias and still get high-quality candidaes, not harder.

For deaf, since one of the features is captioning a speaker.

On the one hand, I know all too well that the AI will screw it up some.

However, if you watch closed captioning, you know that the captions are already frequently messed up, long before even AI was a possible strategy. Usually the live captioned stuff had lower quality, but you'd see it in scripted shows too.

I also wonder about the converse, captioning someone using sign language for those that don't know it.

But that FOV is just so tiny....

Not disagreeing with your argument, but even if all of that could be fixed, fundamentally any anti-cheat that isn't going to be defeated relatively easily needs some sort of privileged access to stop you modifying the game or running other software that interferes with it in some way. That necessarily requires a degree of access to your system that is dangerous, so anti-cheat software will rightly be told where to shove itself by any operating system with a security model worthy of that title.

I don't see the Linux community ever accepting that it's OK to deliberately undermine that security model just for anti-cheat, as a matter of principle. With so many games even at the highest levels already running very well on Linux, I doubt it will ever be a big deal for most Linux users, even keen gamers, to play the 90+% of titles that work and skip the few that insist on more intrusive anti-cheat/DRM measures either.

It sure would be nice to reach a critical mass where the games companies actively catered for that market, though, instead of mostly relying on tech like Proton to make what is essentially a Windows game run OK.

Unfortunately, there is another possible explanation for the emphasis on TPM that is much more sinister. It's possible that Microsoft and its allies are making a concerted effort to lock down desktop clients in the same way that the two major mobile ecosystems are locked down, to kill off general purpose computing and reduce the desktop PC to a machine that can only run approved apps and consume approved content. It already happens with things like banking apps that you can't run if you choose to root your phone to arrange the privacy and security according to your wishes instead of the vendor's or OS developer's. It already happens on open source desktops, where streaming services will deliberately downgrade the quality of the content they serve you when on the same plan you're already paying for they'd serve higher quality streams to approved (read: more DRM-friendly) devices, and where a few games won't run because their anti-cheat software behaves like malware and the free platforms treat it accordingly.

I am worried that we may be entering a make-or-break period for the survival of general purpose computing with the artificial demise of Windows 10. If the slow transition to Windows 11 as people replace their hardware in the coming years means almost everyone ends up running Windows or macOS on desktops and Android or iOS on mobile devices, there won't be enough incentive for developers of apps and creative content to support any other platform, and all the older versions that didn't have as much built-in junk and all the free alternatives will be reduced to irrelevant background noise because they won't support things that users want to do any more. Your own devices will force updates, ads, reboots, AI-driven "help", covert monitoring and telemetry, any other user-hostile junk their true masters wish upon you, and there will be nothing you can do about it.

Governments should be intervening on behalf of their people at this point because the whole system is blatantly anti-competitive and user-hostile, but most of the Western nations are either relying on the absurd valuations in the tech sector to prop up their otherwise precarious economies or watching with envy while their more economically successful allies do that. So our best hope is probably for the legacy platforms to hold out long enough for some free platform(s) to reach critical mass. And frankly, there aren't many realistic paths to get there. Our best hope might be for Valve/Steam to show that many of those Windows 10 boxes in people's homes can now play most of the same games if they shift to Linux and possibly run some of them better than on Windows as well.

Indeed, No True Scotsman would use STL with C++.

The last job I had where I had to use C/C++, we automatically ran an expensive static analysis tool every time we checked in code. I'd estimate that it only found about half of the potential segfaults, and it made up for that by finding twice as many false positives.

ve never seen a software distribution mechanism as careless and sloppy as NPM. Bazillions of dependencies and no signing of packages. [ ... ]

Rust's cargo packaging system is almost exactly the same way. And the last time I looked, Go's packaging was very similar. And package signing won't help if the maintainer's key/cert has been exfiltrated and cracked.

This is what you get when you embrace DLL Hell -- the idea that you should pin your program to a single specific revision of a library, rather than, y'know, doing the engineering work to ensure that, as an app author, you're relying only on documented behavior; and, as a library author, to be responsible for creating backward compatibility for old apps linking to old entry points. Sticking to that principle lets you update shared system libraries with the latest enhancements and bug fixes, while remaining relatively sure none of the old clients will break.

"Sometimes you have to break backward compatibility." Agreed, but the interval between those breaks should be measured in years, not days.