Tesla has switched their newer lines to casting. It seems like it's mostly about manufacturing efficiency rather than strength -- they can cast the whole underbody of the vehicle as a single piece rather than having dozens of stamped pieces that need to be welded or riveted together. Initially they were doing the front half as one piece and the back half as another, but they've needed to get bigger and bigger (higher- and higher-force) machines as they've gone from two-piece to single-piece construction.
The whole point of the current `unsafe` mechanism is that you have to explicitly opt one specific chunk of the code out of stricter checking, and it's obvious which chunk of code it is that needs that behavior (the bit inside the block) and which isn't (everything else) -- that bit can be subjected to extra scrutiny upon review, and there's a community-wide shared understanding of what invariants that chunk needs to uphold for the program as a whole to be sound. I'd argue that the alternative mechanism you propose (a compiler flag), which just opts the whole program out of strict checking, _is_ fundamentally philosophically different. Having the opt-in/opt-out mechanism be block-level is a language-design/philosophical decision.
It seems like for that to work, you'd need to be mostly reliant on day-pass-type customers rather than longer-term ones -- if I were looking for a permanent arrangement to house a small team, say, I'm not sure I'd want to sign with someplace I knew was only going to last until they found a permanent tenant. But I think the day-pass people, in practice, are a pretty small percentage of total revenue.
This is "burst wave lithotripsy," compared to the somewhat-similar existing "shock wave lithotripsy." The new iteration apparently works more quickly and causes less pain, so can be performed while the patient is awake rather than requiring general anesthesia, which would make it practical to do it in, say, an ER or regular doctor's office. See https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10117400/#sec2t... for a comparison of the mechanisms.
I think they're mostly targeting customers who want an AWS- or GCP-like experience from a developer perspective (compute is abstracted and you can provision it with an API, etc.), but want to own their own compute infrastructure and have it on-prem. That market has mostly had to cobble together consumer-inspired HP/Cisco/whatever stuff historically (like, one of the early talks about the Oxide value proposition was complaining about why every server in the rack needs a CD drive, which was the norm from Dell), because the kinds of stripped-down, super-efficient hardware designs the hyperscalers were building weren't available to the general public, so this is that: hyperscaler-like technology for people who want to own it themselves.
I think the motivations for why people would want to own their own are probably a mix of financial (at a certain scale there's a tipping point and it gets cheaper), and regulatory/compliance/whatever, like if it's healthcare data, or defense, etc.
It's more like: you have a machine that evaluates whether the peanut butter you're producing is up to spec (right viscosity, right fat content, whatever), and if your machine complains and says your peanut butter isn't coming out right, you need a way to know if the problem is actually with your peanut butter or if it's that your peanut butter testing machine is miscalibrated. If you put the known-acceptable reference peanut butter through it and your machine says that's fine, it's your peanut butter, but if that also produces an error, you know it's your machine.
> I assume the more I watch the less each creator gets?
I think this is true, but it still works out such that given the amount the average Premium subscriber watches, creators' typical per-view revenue is still higher for Premium views. They can apparently see the breakdown (ad revenue vs premium revenue, etc., ad-supported views vs premium views, etc.) in the backend, and various people who make Youtube content for a living have reported this to be the case. See, e.g., https://twitter.com/hankgreen/status/1513177490730061829 .
That's not the claim. The claim is "because we know smart people have gone crazy, we know being smart and being crazy are not mutually exclusive, so someone being smart isn't disqualified from also being crazy." Which seems obviously true.
> The next question in my mind is whether incorporating private standards into legislation without compensation violates the the takings clause of the constitution.
The standards cost money to produce, so it seems like _someone_ should pay for them, otherwise there'd be no incentive for the standards bodies to produce them in the first place. So I think maybe the best outcome would be a finding that these _are_ a taking, and that then governments who want to incorporate them by reference should have to pay to be able to do so (but then they'd be free to read to the public).
> presumably in this particular case, the companies that wrote these particular fire protection and electrical standards did so because they wanted them to eventually have legal force.
Sure, but the reason they wanted this is at least part because they could earn revenue from selling access to them (and having the monopoly on being able to do so), which they now no longer can. It's not obvious to me how they would otherwise materially benefit.
This tone doesn't feel especially productive. You asked:
> So what you're saying is that we're A-OK as a society with the gatekeeping of critical safety information behind private paywalls and copyright.
but the comment you're responding to explicitly said:
> I agree with the decision and in principle absolutely agree with what you are saying about our laws.
So like... they're obviously not saying what you're accusing them of saying. They're just explaining why this litigation occurred in the first place (and agreeing with the outcome, which seems to be the outcome you also want).
A nice thing about master passwords though is that since you don't have to type them in as often, they can be very long. 95% accuracy probably isn't good enough to reliably reproduce a sentence-length master password, at least if it's only captured once.
I've also seen several different variants for chess on a torus (imagined -- still played on a flat board, but you play as if each edge is connected to the opposite one). Pieces end up having basically the same allowed moves (IIRC the only required change is that a piece making a move where it loops around the board and back to its original position is disallowed), but a different initial setup is required, because given the regular one, both sides start the game already check-mated. For further brain-bending, you can do a Klein bottle instead of a torus.
Again, I feel like there are so many what-ifs that it's hard to say for certain. If it were to turn out that AVs were dramatically less likely to kill people than human-driven cars (seems not to be the case with current technology, but humans suck at driving, so it's not inconceivable), maybe there would be regulatory/legislative disincentives that would make human car ownership harder or more expensive.
> Really? Since when has making using something more convenient reduced ownership of it?
These will be more convenient, but also dramatically more expensive, and _that_ certainly reduces the ownership of things.
> the mean number of cars on the road at any given time would seem like it would only go up
That's orthogonal to the question of level of ownership -- these can both be true simultaneously.
And it may even be the case that the number of cars on the road goes down. The idea would be that if these services are convenient enough for more people to forego owning their own vehicle in favor of using a fleet, the economics on a per-trip basis would change: now, for car owners, owning a car has a high fixed cost and very low marginal cost per trip, so once you've bought the car, registered it, insured it, etc., you're incentivized to use it for as many trips as possible to recoup your investment. Using a fleet service, you would have low/no fixed costs and comparatively higher per-trip costs, meaning that for some trips, it might be cheaper to take transit, bikeshare, etc. A lot would come down to the specifics of pricing, though, so it's hard to predict with any certainty.
Without directly answering your question, it's worth noting that there are also additional optimizations made available by Rust that are not easily accessible in C/C++ (mostly around stronger guarantees the Rust compiler is able to make about aliasing).
Sort of reminds me the Reply All episode "The Case of the Missing Hit": https://gimletmedia.com/shows/reply-all/o2h8bx which involves a similar search for an unidentified song (though in this case, just one that someone remembers having heard, rather than an actual recording).
Yeah, I also liked the idea but apparently have pretty different tastes. I was pretty surprised to see Pericles and Othello almost next to each other in the rankings. Pericles is... not at all good. There's this whole build-up of dramatic tension and then in a sudden, bizarre deus ex machina moment, the plot takes a hard turn when pirates appear out of nowhere and kidnap one of the main characters for no clear reason. One production I saw at least try to lampshade the absurdity by giving the pirates eyepatches and parrots, which almost made it tolerable.
Also, maybe unpopular, but I don't love the Branagh Hamlet. It feels pretty overwrought to me, and I like Helena Bonham Carter's Ophelia better than Kate Winslet's.
Sure, but I'd think that makes the economics of operating fast chargers all the more tenuous. Like, if we assume they're mostly needed for people away from home, that means both that they're essential to adoption of EVs in general (people want to be able to go on road trips) and that they suddenly go from being a thing people need all the time to a thing whose demand has extreme fluctuations seasonally, or with the economy insofar as it affects people's seasonal income, etc., just like other travel-oriented businesses like hotels.
(I should also note: so far the EV adoption story has been much better for people who own their homes and have dedicated parking, but as someone who lives in an urban neighborhood where most car-owners park on the street, I think expending to markets like mine will necessarily mean more public chargers are needed.)
To your last question: the nature of having to attract clicks means youtube creators have to choose reductive titles that tease some particular nugget of information that's going to be discussed, but it seems pretty clear (at least to me) that the intent of this video isn't _just_ to answer that question, and that he expects that his audience will also appreciate a slice-of-life experience of him interacting with the crew and understanding what life is like on the submarine in broader terms than just how they get air.
With respect: if all you wanted was the answer to that question and nothing else, Googling produces it much more quickly than you'd likely ever get from a video, and you might just not be the target audience here. I personally enjoyed the whole series as-is, and it seems from the YouTube comments like many viewers did as well, but obviously YMMV.
They're still pretty squirrelly about the specifics of their commitment, though. In particular, it's not clear if it's hourly accounting (but the fact that they don't say so suggests to me that it isn't), and there doesn't seem to be any discussion of additionality. In other words: it might be that exactly the same amount of renewable generation, and exactly the same amount of fossil-fuel generation, are going to occur with this pledge as would have without, and this is just shuffling some numbers between columns in a spreadsheet.
> the only factor in that transition was the complexity of the nervous system
It seems likely to me that some arrangement of nerves is possible that's comparably complex to ours, but does not produce consciousness. (I dunno, maybe some organism with much more complex sensory organs than ours that devotes so much complexity budget to that that it only has enough left to devote to general cognition to give it the intelligence of a mushroom, who knows). In other words: I suspect complexity is necessary but not sufficient for consciousness to occur. I don't think that takes away from your suggestion that consciousness in AI systems is _possible_, but I don't think it's the case that it's an inevitable outcome if only we can make our systems sufficiently complex. There's probably something about the specific structure of the complex thing we'll need to master as well.
It seems like in the immediate term, and possibly forever, part of the process of decarbonizing the electricity sector is going to be demand management -- we have a finite amount of clean power available, and bottlenecks (interconnection, permitting, etc.) on the speed with which more can be built, so some of the work will just be discouraging use. It doesn't strike me as unreasonable that, if _some_ use needs to be discouraged, the state should have some opinion as to which use it is that gets the short straw, especially since a flat tax on all use would be super regressive: if higher electricity prices mean some retiree in the sun belt can't afford their AC in the summer and dies of heat stroke, avoiding that by shifting the costs onto crypto bros seems like a no-brainer.
All of this probably becomes even more true as load-shifting becomes a bigger deal. Like: currently, replacing gas-powered boilers for 24/7 industrial processes with electrified heating isn't anywhere near cost-effective if you have to pay peak electricity rates for heating, but there are a bunch of new heat-battery companies spinning up that will take advantage of wholesale fluctuations to generate a ton of heat in the middle of the day when the sun's out, heat up a big box of rocks or whatever, and they siphon off the heat over the course of the day as electricity prices spike in the evening. This could dramatically accelerate industrial-heat decarbonization (which is something like 30% of US GHG emissions), but it only works if there's a price trough to take advantage of; if there are crypto-miners there to gobble up all the curtailed power and stabilize the price, the economics don't pencil out, and we're stuck with gas heat for factories. Again, doesn't seem unreasonable as a thing the state would have an opinion about.
In the Spaces thing, he described it as being like an upside-down shower head, and to expect lots of steam billowing out during launch instead of dust, so I don't think it's a closed system.
I think it blowing up or otherwise failing was probably the most likely case and they didn't really hide that fact, but I get the sense that they expected that it would make it a little further along, and not destroy the pad, such that the mess would end up in the ocean instead of on land. This is also littering of a sort, but it's also how every rocket stage pre-SpaceX has met its demise anyway, by design (except for the Chinese ones that seem to have a tendency to come down on land, often in populated areas). Dropping rocket parts in the ocean is just an accepted part of doing business, for better or worse.
The typical approach (a deep flame trench) is likely impossible due to the high water table so close to the beach. Some other novel approaches are likely possible, and I'm sure they'll now pursue one, but coming up with and building a design would probably have taken awhile, and they apparently did some testing that suggested they could get away with not doing it, at least for a bit. It's now obvious that they were wrong, but if they had been right, they would have acquired a bunch of data that would have unblocked a bunch of additional development on the vehicle itself that could have proceeded in parallel with improvements to the launch facilities (the one they're already building in Florida has a conventional flame trench) that they otherwise would have had to wait months for while more-extensive launch facilities were built, so it wasn't necessarily irrational, knowing what they knew at the time.
Even assuming an absolute worst-case scenario climate-wise (no mitigations whatsoever, massive sea-level rise, etc.), Earth will still be dramatically more hospitable to human life than Mars for the entirety of Musk's lifetime, and very likely his children's and grandchildren's lifetimes as well, barring some mass-extinction event like a nuclear war. Even if a colonization effort begins soon, living on Mars is likely to be extremely difficult for centuries -- we can't breathe the air or grow anything in the soil, our biology is likely ill-adapted to the gravity, there is no liquid surface water, etc. -- and the idea that there will come a point in the foreseeable future where it makes sense to escape there because it's easier than being here (because of pollution or whatever other reason) is difficult to fathom.