Once RFK Jr is out I genuinely would not be surprised to see Holmes get a pardon and HHS nomination.

Many smart people, very smart, have been saying she's been treated very badly by the Biden and the Dumocrats, very badly indeed, it's horrible what they do to people like her. The other day, a very big man, huge muscles, big supporter, he came up to me and he said with tears in his eyes, Sir, he said, I was very sick, and he told me but I shouldn't say with what, but do you know what he did? He took just a drop of his blood, just one, teeny drop, and he gave it to Thanos and bip bop bip! Testicular cancer totally gone! And now they put her in jail, Elizabeth they call her. Holmes, like a certain, detective, maybe you've heard of Sherlock? And Mr. Watson. Like Sherlock Holmes and Watson. Dr. Watson. Very famous detectives, my uncle knew them he said, they solved crimes, many, many crimes and everyone said couldn't ever be solved by anyone. He found Jack the Ripper and made him stop his crime spree. Slashing dead women and blood everywhere. People say he saved New York, and I believe it. Sherlock Holmes. So beautiful, so smart, beautiful, blue eyes, some people say she looks like Ivanka. And it's horrible conditions, they don't give her anything to eat, almost nothing at all, and she'll be there, in solitary, for 50, 60, maybe 90 years. Some people are saying maybe it's forever. She won't look like anything when she gets out, and the drops of blood, they can't save anyone. It's so unfair what they do to people like her.

Sigh. that's 80um 150um except imagine it's a micro symbol instead of a normal u

Thanks, slashdot's html sanitizer.

I think that if you want to generate electricity from sunlight, this is probably strictly worse than PV cells.

They're effectively extremely thin aluminum foil that has been tooled using a femtosecond laser and mounted to some kind of substrate, so the materials are dead cheap. Femtosecond lasers are somewhat uncommon compared to other industrial lasers but not particularly exotic these days; much (most?) LASIK surgery uses them now.

Throughput is a challenge, though -- the lasers remove very small amounts of material, so they take a while to do their thing. But we've proven ourselves capable of getting awfully high throughputs when making materials with small features, so that's probably solvable. (And these features aren't particularly small -- the grooves are between 80m 150m wide.

Harvesting the salt might be a challenge, but you're basically scraping salt off of a flat surface so that should probably also be quite feasible.

Musk also could have just filed suit within the allowed time, which is the correct way of doing this. It's not like he didn't know about OpenAI's actions; none of this was secret and he clearly knows how to file a lawsuit.

The actual goal here was likely to make OpenAI go through discovery, waste a bunch of money, and get company officers up on the stand to say embarrassing things. Which they absolutely did.

It's not entirely clear to me why Altman didn't have the suit dismissed way back in the day; the statute of limitations question was not, like, a challenging one and the judge even basically said "yeah we could have dismissed this ages ago." Unless, of course, Altman wanted to make Musk go through discovery, waste money, and get on the stand to say embarrassing things. Which also happened.

Congratulations to both parties, I guess.

It's not a "not guilty" verdict, it's a procedural verdict; the statute of limitations had expired prior to Musk filing suit. (Also, civil trials don't find guilt in the way criminal trials do.) You can appeal that ("the court calculated the dates wrong" I guess?) but generally you will fail because if you could show the court calculated the dates wrong, you would not have gotten tossed in the first place.

The exception would be if you had really strong evidence the court was in the bag for the opposition and was treating you unfairly, but AFAIK there's no evidence of that here.

This said, Musk has absolutely bottomless pools of a) money and b) resentment so God knows he'll probably appeal just to be a dick. I'm not sure if this is a "he will have to pay OpenAI's lawyers and court fees" situation but even if he does, it's still probably more annoyance than the money is worth.

From the article:

In a unanimous verdict, the jury in Oakland, California, federal court said Musk had brought his case too late. The jury deliberated less than two hours.

Eleven days of testimony to discover the statute of limitations had expired, which should have been trivial to calculate? And, as far as I can tell, the judge warned them about well in advance? My God. What a colossal waste of everyone's time. I hope court fees were hefty indeed.

I totally get this, and also we have quite a lot of experience with cranial ultrasound; we've been using it for imaging in newborns since the 1970s.

So yeah: We are indeed testing a new thing but we do have a tremendous amount of data on the effect of ultrasound, at various intensities, on human tissues.

Also, it was radium that made paint glow, not uranium.

Uranium was used in ceramic glazes (Fiesta's orange was the most famous example); despite widespread worry about "omg it's radioactive" the dose you'd receive from using those dishes for your food is less than a background dose, and much lower than you get if your house has radon. In practical terms, it's very safe.

Yes, still probably don't eat off them; keeping your dose "as low as reasonably achievable" does entail not getting any dose from a pretty orange bowl if you can just use a different bowl.

Wait; you're criticizing someone's argument from authority and your rebuttal is effectively that you thought the movie was boring?

I am not saying the Oscar committee is incredible or anything but I don't see how your opinion is any more valid. You can't just beat argument from authority with argument from non-authority

For what it's worth: All ballistic missiles are hypersonic. They've fit into shipping containers for yonks; Iran publicly launched some from a container back in 2024. Some of them have some limited maneuverability at the terminal phase to improve targeting / defeat defense. It looks like this one does.

I mean the real answer is that Lockheed Martin's missiles are for a different purpose than these are. The US cares (well cared, at least; God knows how much the current leadership cares) very very much about reliability and accuracy, because our goals have been to hit small, tactically important targets while avoiding civilians. Hitting a school full of kids is bad, actually! Like, really, really bad! We try very hard not to do that! So we pay $3.5M for a Precision Strike Missile or $1.7M for an ATACMS, because they can accurately, reliably hit a targeted building from hundreds of km away. Paying less for the missile also doesn't help you as much as you might hope; getting good intelligence and targeting data (again: hitting schools is bad) is also fantastically expensive.

Accuracy and reliability may not be the primary goals for a $100k missile. If this weapon is intended, instead, to provide a credible deterrent, you don't need accuracy or reliability. You need numbers large enough to make defending against the missiles impractical, and make the cost of fighting intolerable to potential adversaries. You also can't defend with cheap interceptors. Incoming missiles can be (almost) anywhere in 3D space. Defenses need to detect, acquire, and intercept targets with incredible precision on ridiculously short time scales.

And while $100k is cheap, it's not _that_ cheap -- retail for a Scud might be $500k-$1M. Cutting production cost by a factor of ten isn't so unreasonable with a ground-up design and modern materials and tooling.

I am also not at all convinced that this missile, in actual practice, will be particularly inaccurate or unreliable. Chinese companies are extremely good at mass-producing precision devices out of challenging materials at absurdly low cost. Time will tell, unfortunately.

The USG and allies have traditionally been very willing to pay top dollar for missiles. Lockheed has no reason at all to make a cheaper variant of the PSM; we're happy to pay the full rate. For many reasons, I do think the US should develop the capability to build precision stuff in larger quantity at lower cost. The military isn't likely to drive that, though.

But: We're never going to try to acquire the cheapest possible means to deliver bombs to general areas. There just aren't many (any?) situations where the ability to blanket an adversary's city with missiles actually helps us achieve our goals.

Anyhow, I am hopeful that maybe, someday, folk will decide that missile defense is a fool's game and that maybe leaning on diplomacy a bit harder is more effective at solving conflicts.

Yep: The fundamental problem here is that the US has spent a long time with the values "housing is a luxury good; home ownership is an investment." In my opinion, a more appropriate set of values is "housing is an absolute necessity; home ownership is a luxury good."

The funny thing is that even with current very owner-friendly policy, home ownership is not actually a particularly good investment.

The funniest thing is that I don't think I have ever read a "coffee is bad for you" article. I've read some "coffee, with or without caffeine, is associated with lower liver disease" articles and some more recent "coffee and tea with caffeine is associated with a lower chance of dementia" ones. I guess large amounts of caffeine can cause health problems, as my gastric ulcer a few decades ago can attest to. But... yeah. That doesn't strike me as inconsistent.

I guess there's also the "coffee drinks with sugar have the same negative health effects as other sources of sugar" but that's all "lots of sugar is at least kind of bad for you" which has been pretty consistent advice for a very long time.

Don't get me wrong: Nutritional science is generally a tire fire. Getting people to change their diet is extremely difficult, retrospective reporting on food intake (content and quantity) is very unreliable, effects may take years to be apparent, and dietary changes tend to be correlated with other things that we also expect to affect health.

But coffee itself? Pretty consistently somewhere between "not harmful" and "moderately good for some aspects of health" as far as I have read. Maybe GP was thinking about the "1-2 glasses of red wine is good for you" thing from a few decades ago?

And yes: Science communication could do better, and also... for whatever reason, society seems to periodically glom on to random weakly-supported nutrition fads and grifters publicize the crap out of them to make a buck before fading into the background.

I think the problems Tailwind is facing are a perfect confluence of factors. CSS in general has gotten a lot better over the last several years; you need way less detailed knowledge to implement a visual design than you used to. The CSS spec is also very well-documented, and there's basically infinity CSS out there for models to train on.

Tailwind's value proposition is that they make it easier to implement a consistent-looking visual style without writing a bunch of CSS; in particular, they handle the tricky parts of layout and save you a lot of typing.

But if CSS is good enough that the tricky parts of layout aren't really that tricky, and AI agents will handle the typing for you... the role Tailwind fills is just way less important in the web development ecosystem. Tailwind is great! It's also a big dependency!

Similarly: I think AI will also potentially make a huge difference in the javascript ecosystem. If the AI can implement a bunch of the libraries you've been importing, your dependency tree gets less complex.

On one hand, this is terrible for a lot of open-source projects. Who's going to support all these little projects that see declining use?

On the other hand... business needs and models change. The need for people who can do visual and UX design, taking platform and actual use into account, and turning that into a good, maintainable implementation, won't go away. The CSS programming there isn't the hard part; the challenge is understanding the larger context and managing the project and delivering the technical parts of the project quickly, and adapting it quickly, are.

Can you provide a great open-source CSS library that you use in your projects, essentially as an advertisement for your design services, and insurance for if your design firm goes away? I would guess yes!

Can you provide paid access to templates that use your framework? I... I'm not so sure about that. Maybe a little? But I think that knowing what design choices you want to make for a particular project and having the judgment to find potential good fits and take feedback usefully and iterate are much more valuable skills.

It's a different business model. Maybe not a fun one to pursue. But you have to adapt.

China is perfectly capable of making products with excellent quality control and safety; they're as good or better than anyone else. They manufacture iPhones for Pete's sake.

Yes, China makes a bunch of the cheapest, most terrible products out there, because a lot of people want to buy cheap stuff and Chinese manufacturers want to sell cheap stuff. Sometimes, that means they're buying something unsafe. Going to the bargain bin for batteries is not a great plan.

Would I buy a BYD on the first year of import and park it in my garage? Probably not. But I'd read the reviews and watch a few teardown videos and if they're competently manufactured, I don't know why I'd be any more worried about one than a car by Ford, Nissan, or VW.

That's the thermal battery. The reactor core itself is cooled by liquid sodium metal. Sodium is a nice reactor coolant because it only produces one short-lived isotope (which in turn decays to a stable isotope) but it is still literally liquid sodium metal so keeping it dry is essential. Designs that run the sodium metal loop through a water loop to drive a turbine are inherently pretty spicy.

The Natrium design puts a sodium salt as an intermediary -- the sodium metal dumps heat to sodium nitrate, sodium nitrate dumps heat to water, spins the generator.

This is nice because it means you can run the reactor at its highest efficiency all the time and use the salt as a buffer to vary electrical output. It also means you don't have water and sodium adjacent to each other -- you can actually put them quite far from each other.

Yes, corrosion of the reactor coolant system in particular seems like it might be a major challenge. Material science has gotten better, so maybe it's okay now? Anyhow, it's hard to test without building the thing.

This is one reason they make flexible adapter cables instead of just rigid dongles.