It is a false equivalence. Other than gambling, everything on that list is something physical that causes physical harm. Gambling is psychologically addictive for some people, and can cause psychological harm. Free speech is neither physical nor psychologically addictive. Lumping free speech in with gambling, alcohol, smoking, or dangerous toys is pathologically dubious to the point of being insulting to anyone with common sense.

But the only way you can break the "think of the children" argument is by undercutting the claims that free speech harms children in similar ways.

I never said it did. I said it gave a false equivalence between protecting children from free speech and protecting children from dangerous toys, dangerous substances, and environments that feed addictions.

No my argument is that you can't win by pointing out that they're playing on emotions, because it is precisely that playing on emotions that shuts people's brains down, and it won't come back up just because you point out that they're playing on someone's emotions, because nobody likes to admit that they got played, so they will double down by default rather than feel stupid by recognizing the truth in your words. This is a mistake that a lot of politicians make, and it usually results in them losing.

The only way you can win is by making it pointing out how stupid the comparison itself is. By doing that, you cut the knees out from under the emotional part argument, so that the reader doesn't feel stupid, but rather a bit betrayed. And you give the reader the opportunity to feel smarter than the original author for recognizing the truth in how bad the original author's argument is, and now the emotional argument no longer carries any weight.

That or counter with a stronger think-of-the-children argument that is pro-free-speech.

I don't focus on that point. I just recognize that you're not going to stop people from screaming "Think of the children." They're going to do that no matter what, because they know that people become completely irrational whenever kids are involved, and they know that using kids as an excuse will get you past a large percentage of people's bulls**t filters.

The only thing you can really do is destroy the fundamental arguments that they use to support their position that doing this specific thing for children is useful or good, and that false equivalence is the foundation upon which that claim is built.

I mean, their government's whole point was trying to draw a false equivalence between selling physical goods (importing a product into a country) and running a website (allowing free speech to be served by a machine that provides content when asked). Tearing down that false equivalence between taking an action to bring something into a country (active, requiring actions by someone in that country) and merely passively having your speech available to someone in a country really is the crux of any rational argument on the subject, though I'd have used a lot more words.

Depends on what you're trying to prevent. If your goal is just to prevent the government from misusing location information in ways that could lead to the wrong person being arrested for a crime because of coincidence, you can solve that by simply passing a law that declares all location information gathered in any manner other than direct surveillance by law enforcement to be hearsay, and thus inadmissible in court. That would still allow limited use, so long as you end up with enough evidence without that data, and so long as you had enough other evidence to show probable cause before obtaining any warrants to get the other evidence, but it would mostly destroy the usefulness of third-party location data to law enforcement.

The line is "user behavior". Lowering price based on demand and competition are fine. As soon as you drag user behavior in, one of two things will happen:

• Users will figure out how to game their behavior so that you will give them a lower price.
• You'll accidentally do something that causes illegal discrimination against someone based on a protected characteristic and find yourself fined orders of magnitude more than you gained.

You may do both at the same time. This sort of behavior, IMO, falls under the category of "suicidally short-sighted business decisions", and is something that nobody in their right minds should touch with a ten-meter pole.

Tiny. They're a startup, and they've only been around since 2023.

That would be Elon Musk. He thinks the law doesn't apply to him because he has money. Sadly, he is probably correct.

Their belief was that because they were temporary turbines on trucks, they were exempt, in much the same way that buying ten thousand gasoline generators would arguably be exempt. The courts concluded otherwise. Musk cuts corners and skirts the law. Always has, and probably always will until he ends up in jail for it.

But the existing laws worked (eventually), and there's no reason to believe that having more laws would make him more likely to not break the law.

I know 80,000 households sounds like a lot, but that's only because households use approximately no electricity compared with industrial power users. To me, this reads like someone sensationalizing a non-story to get more views.

150 megawatts is one moderately large gas turbine. It's 60 Tesla supercharger stations. It's somewhere around 5 to 10 percent of what Ford uses for manufacturing in Detroit alone. This is not a huge amount of power. The largest gas turbine in the world produces 840 MW. You could almost put six of those data centers on that one turbine.

Not that it is okay to burn natural gas for data centers, mind you, but the problem is that there are inadequate laws mandating clean energy for data centers. It shouldn't matter how much power or water you use, so long as you produce it in a way that is environmentally responsible.

It's also worth pointing out once again that you can't conserve your way to clean air. The most expensive power will always be nuclear, so when you cut production, you'll almost always be cutting clean power production and making the grid less green on average. And reducing power consumption causes a reduction in power plant construction, and the power plants that are being built now are cleaner and more efficient than existing plants, with a greater focus on solar, wind, and battery storage, so by reducing power consumption, you're actually reducing the incentive to make the grid greener. You'd have to conserve almost all the way to zero power consumption before you would actually have a positive impact when it comes to the energy mix, and that's a really bad idea.

Is that more or less than the recharge rate? Like I said, companies use water in places where there is a lot of water.

Sure it would. Most of the big companies building data centers do exactly that. It's only the small startups that use ridiculous amounts of grid power and try to find a place that's willing to connect them. The big companies provide their own generation so that they can hook up wherever they want, because while it might save money to rely on the grid, it wastes months or years trying to get the permits, and every extra month puts them at a disadvantage. And a lot of the data centers being built are pretty green, too.

This isn't about the strain on the grid. If it were, the grid operators would be screaming.

And this isn't about driving up power rates, because that's not how it works. It would drive up power rates whether they build in Ohio or any neighboring state, because entire regions share a grid, so if you're not adding more generation to the grid, you're causing someone to buy power from generators that cost more and/or charge more.

No, this is and always has been about luddites looking for excuses to stop progress. It's just like the trolls who scream half-truths about water usage, ignoring that the data centers that use lots of water do so because they *have* a lot of water to spare at that location, and use different techniques when that isn't the case.

The kinds of people who come up with these idiotic laws don't care about facts. They only care about screaming loudly and hoping people will believe them and do something about AI or "big tech" or whatever the shibboleth is this week. They want to do something, and this is something, so they must do this. They want to "stop big tech", without any real logic behind it beyond that it's big tech, and therefore must be bad. You get the idea.

If they limit the consumption per data center to 25 MW, they'll just spread it across multiple data centers. While there are some advantages to doing it all in one building (lower electricity waste, lower water use, lower environmental impact, slightly lower cost), it isn't as though that's a showstopper.

Besides, it's easy to build a data center with net power consumption less than 25 MW. You just throw batteries at the problem, and build a massive solar farm on-site.

To me, this is basically workers throwing their wooden shoes into the machines hoping that it will prevent industrialization, and it will be exactly as effective. If you want to pass laws limiting the negative impact of AI, pass laws about that. Trying to stop job loss from AI by blocking data centers just means that they'll be built somewhere else, or tweaked to barely conform to whatever laws you throw at them.

They're buying access to the data from someone else, so not necessarily.

Yes and no. Yes, getting real-time access to data is possible. No, that doesn't mean that you'll save money with some vibe-coded tool.

Bloomberg is doubtless making a profit at $24k to $27k per user per year, but don't expect the cost to drop to free or evey cheap if fast access to real-time stock data is a hard requirement. Most of what you're paying for is the delivery cost for the data, not the software or the hardware. Specifically, for larger installations (more than about 4 terminals), if I understand correctly, you're getting dedicated bandwidth to the firehose (leased line).

And of course, it goes without saying that using AI to analyze that firehose is not realistically possible. There's not enough TPU/GPU capacity on the planet. And I'm assuming that's what people are really hoping for. :-)

s/cheap/expensive/

Sure. And that's why it is so important to establish ahead of time what we consider to be okay or not okay.

Always was a boutique thing. It isn't the voice that makes acting meaningful. It is the inflection, the pacing, etc., none of which AI is very good at (or at least it wasn't good at it the last time I checked). There's just too much contextual understanding required to know whether "I understand" should be read in a militaristic "Yes, sir" way, in an downtrodden "I just got criticized fairly" way, in an "I'm trying to convince you that what you're doing is wrong" way (usually followed by "but"), or in a caring "I know what you're going through" way. The amount of prompting to get the right results exceeds the effort of having a real person go through the script and act out the lines. And even with using the voice of a dead actor or actress, that's still true.

I would like to see these credited as "[Living Actor] in the voice of [Dead Actor]". That way, the voice actor gets credit for the acting, and the original actor's voice gets credit for the sound.

Yes. It's a classic example of the slippery slope fallacy. It presumes that no one is capable of drawing a line between "coming for the communists", which could be as little as canceling members of or banning the existence of a political party that has harmed a particular country or region and "coming for the Jews", which is assumed from context to mean racially or religiously motivated genocide.

Slippery slope arguments are generally fallacious, not because you can't see how one thing being normalized can lead to something worse becoming acceptable — this is, of course, closely related to the concept of an Overton Window — but because it ignores the existence of free will, and the existence of people actually having moral boundaries that they won't cross. It ignores the existence of free speech, and people standing up when others' behavior crosses those lines and saying, "This isn't right." And so on.

It is never necessary or useful to prevent reasonable behavior because of slippery slope arguments under the premise that someone coming to accept those reasonable behaviors could come to accept some unreasonable behaviors that are somehow related. In fact, doing so actively weakens the public's sense of morality by making people think, "I thought this was okay ten years ago, but people were protesting against it. Maybe these things that those same people are protesting against now are also okay." It polarizes people unnecessarily when it should be unifying people. In short, it is why our politics in the U.S. are as screwed up as they are now.

Negligibly reducing opportunities, at best. It's not like an extension of a long-dead series with most of the original cast is something that happens frequently. This is the 0.001% case, if that. So .001% of shows get to do something that seems almost miraculous (acting by dead people), because their circumstances are exceptional.

A bigger concern is whether that voice cloning technology then leads to a few famous actors and actresses selling their voices and destroying the market for actual voice acting. But that does not inherently follow from this specific use case. Most of those folks wouldn't want their voices to be associated with projects without their sign-off, which is contrary to the desires of the folks who would want to use AI in place of live talent, and there's no reason to believe that they can find middle ground.

And this all assumes that the AI approach eliminates the need for a voice actor. Realistically, even with AI, there will probably be an actual voice actor saying the lines initially, and the AI system will then match the inflection of the actual voice actor, but in the voice of the original cast member. So chances are, at least for the time being, you'll see both the original actor's family *and* some new actor getting paid; the only differences between that and your approach are how it sounds and how it gets credited.

In other words, much ado about nothing, IMO. The real problem is the step after this one and/or the step after that.

Because this is a continuation of an existing show and the original actor is deceased, if done properly, it could fall solidly on the "okay" side of my moral/ethical bright line rule.

• Using AI as a way of honoring the memory of a fallen member of the cast, with the permission of the family, and while paying the person's heirs, is okay.
• Using AI for the purpose of not having to pay actual actors is bad.

The area in between those two extremes is where it gets grey, e.g. if they have the family's permission, but don't pay them anything, or if they don't have the family's permission. In those cases, I would err on the side of calling it dubious, but others may disagree, depending on the details.

No, that was math.

$17 per hour * 8 hours = $136. $136 / $10 per gallon is ~13 gallons. This is, of course, an overestimate because of taxes, but the point is that even at $10 a gallon, you're realistically an order of magnitude away from fuel prices eating your entire paycheck for most non-menial work.

At $10 a gallon, even working at Burger King, you'd have to have a round-trip commute of probably 30-ish miles before you'd lose your first hour's paycheck to fuel. And most people who work at Burger King live relatively nearby (ignoring weird edge cases like the Bay Area), because at that skill level, you're not likely to need to drive a long distance to find a job.

Ironically, it is the cost, and more specifically, the low profit margin, that likely has driven these improvements. Even with a low failure rate, every extra minute spent doing repairs cuts into that margin. While you could go full disposable, that's a bit more objectionable in a laptop than in a cell phone.