This is a molecular dynamics (with a specific TIP4 water model) and AI. I would characterize this paper (and basically most MD/AI simulation) as "guided hypothesis generation". I am skeptical that this result will hold up to experimental validation. It is incredibly easy to generate molecular dynamics results that look perfectly reasonable but have nothing to do with "reality" (i.e., how a lab experiment with the same system would turn out.)
One of the biggest challenges with water is modeling proton transport (i.e. pH, as water is a weird self-ionizing material.) Protons move too fast for MD steps to be stable, so they have to be approximated in some way. That is one reason why there are so many water models to choose from. Each model has a trade-off that is fine in some contexts, disastrous in others.
That's a quote from Corey Doctorow, not Sarah Wynn-Williams. I read her book. She was pretty careful to use your language (i.e., that it was offered, but not implemented, and was China-only data from what she related. Not that that is great either, of course...)
Her main allegations (that Facebook/Meta optimizes for profit at the expense of everything else) seem pretty unsurprising. I mean, given what has been observed, is this in any way controversial?
The intuition is that in order to apply a force through some distance, I have to change the velocity of an object by dv. But, the distance I just traveled also depends on the current velocity v. That's why the total energy available isn't just simply proportional to velocity - every time we change v, the amount of force available goes down, too.
Summing all the little bits of energy dE over our velocity changes dv, from the starting velocity down to zero, and we get the formula for kinetic energy.
BTW, the intuition here really starts from the idea that force = momentum change with time. The definition of "force", "momentum", and "energy" can be maddeningly circular, even if we have clear mathematical representations and a common world we experience.
>All I do is read books, paint, fix-up the house, tend to the garden, play boardgames, and generally do things that require me to be present when at home with my wife and child, rather than in someone else's version of the world
As "cool" as this vehicle is, I have a hard time seeing it make much progress in the market beyond a novelty. It's basically a covered 3-wheel motorcycle, and that rear wheel is going to hit every piece of road debris that 4-wheel vehicles push to the center of the road. Commuting on freeways in the US would be dangerous. And as a 2-seater it would mostly appeal to the sports car crowd, but it simply isn't a sports car.
It's somewhere between an e-bike and a real EV, so it would need to be priced there too ($10-20k-ish) on a low-margin product. But it will have very narrow appeal. Hard to see how the business makes sense, at least in the US. And the rest of the world has BYD cars.
"The digital commons of 2026 is defined by its grifters. So it's not purely tech's fault that its now seen as a sort of avenue for getting rich quick and amorally"
I am not so sure I agree with this take. The "nerds" are building incredibly powerful technologies (Amazon, Starlink/SpaceX, search, algorithmic social media, AI, etc.) that literally control our lives now. It isn't any great mystery that the tech titans realized they had this power, and hence are questioning whether democracy is some outdated concept. They all want to be Plato's philosopher (or in this case, technologist) kings. At the risk of sounding like an AI, it isn't just grifting (or a con game) - these guys really do think of themselves as the new feudal lords. So I don't think this author is thinking big enough...
To quote good old Chester Karass (https://www.amazon.com/Business-Life-Dont-Deserve-Negotiate/...): you get what you negotiate. In an ideal world, people are compensated according to their contributions (and when hired, expected contributions.) This isn't a once-and-done thing, though - compensation (in the form of company ownership) gets adjusted all the time. The flat 10% is just a starting point (a typical CEO level of ownership in a start up, although this number varies quite a lot, due to aforementioned negotiation.) If a CEO screws up badly enough, they might get fired even before their stock vests.
Of course - at founding, if $20M goes into the company at $1 per share, and the CEO gets a 10% equity stake (usually subject to restrictions), then the CEO has $2M on paper (or will have after possible vesting.) Real money in this case came from the original investors that flowed into the company in exchange for ownership, but the CEO can't really do anything with his shares yet. At this point the original investors are taking a huge risk with their money - chances are, they just lost $20M dollars, and probably even more, as it can often take a long time of putting good money after bad.
Once a company starts operating, but before revenue (and hopefully eventual profitability), the valuation is trickier. The share price _should_ be the number of shares divided by the sum of all future profit (minus current debt.) Which is hilarious of course, because no one actually knows the denominator.
That original $2M equity stake can grow to billions if the company ends up making something that a lot of people want or need, so the sum of all future profit is large. Or, much more likely, it will be worth nothing, or a modest amount.
Graham's essay kind of avoids the point of whether ownership of a vastly appreciating asset is "fair", if a bunch of other people help that asset to appreciate.
He was talking about an equity stake in a start-up. Although on paper it is worth $2M, it is (probably) not liquid (i.e, the shares can't be traded easily, maybe at all.) The vast majority of founders don't literally spend $2M from their checking account to purchase their position in a start-up - they get some ownership as part of taking the start-up risk.
That is the drawback - sometimes important calls come from someone not in my contacts (like an emergency where someone borrowed a phone, or a contractor trying to call you.) Still, the beauty of voice mail is that, if the caller is really trying to reach you, they will leave a message. Some random number with no message is almost assuredly spam. That model wouldn't work with email (you would need an equivalent real-time notification of an email going to a spam folder, then the user would have to decide to send to spam or not send anything.)
Anyone without caller id is also suspicious. Emails have a sender, but it is also about as reliable as a caller id (i.e. not very) when it comes to identity.
This is basically where I (and I imagine many others) have landed with the telephone. Anyone not in my contacts goes to voice mail. Made my phone usable again.
OK we are really nerding out here but I just can't resist :)
The propelled blade weapon doesn't need a control system in the "shape of a human mind" to be effective. A very simple control system would work. If they have ornithopters, then they have control systems. The blades could even work like hawk talons (so an ornithopter drone falling on a shielded fighter, then springing what is effectively a pointed metal trap.)
But more importantly, you are right that if the weapon moves too slowly, then you could just get out of the way. That's true of any weapon, no matter how it is propelled. That's why I said above you have to be able to penetrate a shield faster than a human reaction time, otherwise no weapon would be effective.
True for human generated results too - how many "proofs" of Fermat's Last Theorem were proposed then discarded when a mistake was found, before Wiles?
2. Lack of proper attribution and violation of copyright
I have trouble understanding the difference between a human reading and digesting copyrighted materials (i.e. "training") and AI training. As long as either case doesn't reproduce this material word-for-word without attribution (which is why we have a word for this - plagiarism.) OK, maybe an electronic thinking tool makes plagiarism easier if you are careless. A power saw is more dangerous than a hand saw, too.
3. Dependence and inequality
Any fancy automation tool brings this risk. Is a Matlab subscription or access to a data center "inequitable"? And yeah, if you live in a place where you need a car, then you have a mobility dependence. You can still walk, but it isn't practical.
4. Overhyping of results
I don't see a difference between human or AI generated here, although the newness of AI does lead to people being more interested at the moment. The shine will wear off eventually.
5. Loss of autonomy
I just don't understand why AI is a problem here - the "autonomy" here is about who sets what topics are interesting in mathematics. Is it because every amateur now has a new powerful tool that once only existed in the brains of a few adepts? So they (or AI itself) might have input into what problems are important?
So overall I think this isn't a very effective list of risks. Everyone is nervous that the unique thinking abilities that set mathematicians apart (and I completely get this as a physicist) are being eroded. Until the reliability of the output of AI systems improves by several nines, we need humans in the loop to make sure the output is correct (Point 1.) But given that AI is improving rapidly, extrapolation to the day where humans are effectively out of business is now something that needs to be considered. Just declaring we need humans to do mathematics research isn't going to be enough if that day ever arrives.
>What good are drones, except as weapons of assassination? Can't use them in an conventional war, there are shields.
Presumably a drone could get close enough to push a blade through the shield (maybe use a rocket booster or some other advanced "suspensor" tech) at low-ish speed (a knife would have to travel through the shield on the order of a human reaction time, otherwise even knives wouldn't work. That's about 0.1s/0.1 m, so 1 m/s - ish.)
The Bene Gesserit tried to solve alignment through their breeding schedule, but Lady Jessica threw a wrench into it by bearing a son (Paul) out of love for Duke Leto. Paul wasn't aligned with the Bene Gesserit, and hilarity ensued.
So I agree, I don't see why thinking machines are a worse alignment problem than enhanced humans. See also the Wrath of Khan.
I try not to think too deeply about it, or I won't enjoy the story. How could a society have gravity suspensors, "shields", and lasguns, but no advanced computing capability somewhere? And why use swords/knives for fighting, when projectile weapons are so much more deadly? I still love the books, but I have friends who can't enjoy the Dune world due to its many contradictions.
Edit: I realize that "shields" supposedly made ordinary projectile weapons ineffective - but they had advanced drone technology (e.g. the assassin drone used against Paul, which was also stupid)
Sure - having raised kids myself, and never once hitting them, there are lots of ways to memorably punish a kid. First and foremost, kids really respond when their parents get upset. They remembered when I was visibly angry with something they did. Throw in some "I am so disappointed with you" from the other parent, maybe grounding them or taking something away they want, and we have something a kid would like to avoid in the future. Like any punishment, it can't be used too much or it loses effectiveness (so if you blow up at your kids constantly all you are doing is funding a future therapist's income...)
FWIW, teachers in my (basically all-white) school in the Northeast US in the 70's and 80's would occasionally paddle kids for egregious behavior (like doing something dangerous or open insubordination.) In my low-N observations, it had dubious effectiveness. Kids who were generally good but just messed up once appeared to be "scared straight", although of course I can't say if it was more effective than non-corporal punishment. The sudden shock and embarrassment of being paddled seemed to be effective though. If the kid was a "behavior problem" paddling wouldn't fix the underlying issues.
Also, if you were punished at school, you could expect to be punished at home, if your parents ever found out. Most parents back then would immediately side with the teachers; that's something else that has really shifted in the last half-century. Today, parents are involved to the point of interference, at least in the more affluent areas.
>As you point out, if they designed this thing in the late 1970s, there is no >reason for those giant arrays of drill holes.
Yeah, this was precisely what my original question was about. Weird claims in this thread - we had multilayer boards with CAD layout in 1980, but can't drill through-vias in anything but a complete grid pattern? These are low-volume production boards too, trivial to drill an arbitrary pattern. (Edit: Obviously we did have multilayer CAD designed boards in 1980...)
My own (tongue-in-cheek) guess: someone prototyped this circuit on a perf-board with point-to-point soldering, then handed it off to a too-literally-minded junior engineer to do the PCB layout :) Unless there is some aerospace thing I am missing for having the full hole grid available...maybe they thought it was a weight savings?
I'm in my sixties and reflect sometimes on how much freedom I had as a kid, and why things have changed so much in terms of risks parents are willing to accept.
One correlation with "safetyism" this article doesn't mention: the rise of the two income household (https://www.pewresearch.org/social-trends/2014/04/08/after-d... for the US; the UK appears to be similar.) In reality when we kids were running wild about the town, someone was watching us out their windows. If we got into (or more likely caused :) ) a problem, adults, usually a housewife, would show up quickly from somewhere. Even when we were off in the woods there was a sense that we could find a house where a grown-up would help us if needed (like if some kid's little brother ruptured his spleen on a dare, which actually happened.)
Nobody would call Child Protective Services - you knew it was little Billy who threw that rock that hit Jimmy, so-and-so's kid. You would tell Billy's dad, who would make sure he didn't ever do _that_ again, and that would be the end of it. Now I imagine police and lawyers would be involved. It seems we don't have the informal social connections any more, which were largely driven by someone just being around.
The above link BTW shows that "only" 50% of mom's were stay-at-home in the 1970's. In my specific time and place, many of the moms who did work outside the home had jobs that revolved around the school schedule (i.e., working at the school, or some work schedule that allowed them to be home when the kids were not in school.) The ones with full time jobs like my single mother, supporting three kids through full-time work, were a rarity back then. Maybe my brothers and I had excessive freedom because there simply wasn't anyone to watch over us - fortunately we all turned out more or less OK :)
One of the biggest challenges with water is modeling proton transport (i.e. pH, as water is a weird self-ionizing material.) Protons move too fast for MD steps to be stable, so they have to be approximated in some way. That is one reason why there are so many water models to choose from. Each model has a trade-off that is fine in some contexts, disastrous in others.
My money is on a TIP4 artifact.