I think we are in agreement on most of this, but you don't seem able to see it. We agree that if we stick with the status quo for regulatory involvement that both nuclear is too expensive, and we shouldn't just shut off solar.

I'm not certain if we disagree on whether solar is really a long term option. We may not.

But again, that's all relative to the status quo. I'm arguing that the status quo is not in line with current technology, and until it is, we can't really know that the other options aren't financially viable. It seems to me that they have a high chance of being so. Full stop.

I live 30 miles from a nuclear plant that's been there most of my life. No meltdowns to my knowledge, and Meta just cut a deal to extend its life and expand its capacity, presumably because it's way easier to do that than build a new one. The financial argument is there. There's discussion now, because that plant was originally designed for two reactors but only implemented on (because of additional regulation/licensing cost per my discussions with people there I know,) that they may do that very thing.

There are a handful of NatGas Peakers in the 150 miles from my house. Never heard of those blowing up. Not to say it can't happen, but it's not common.

None of this considers the newer technologies out there. Gen4 Nuclear plants are both cheaper to operate and safer by design. You can argue that we don't have them in the U.S. so "we don't know that," but that's mostly because we haven't been allowed to build them in the U.S., so it's a circular argument.

I mean, I can poop in my driveway. That doesn't mean that's the best place for it.

On the one hand, Solar is currently the fastest energy source to bring on line. Relative to (from what I've read) ~3 years for NatGas and even more for other options, Solar has a quick turn-around.

That said, NatGas and dare I say Nuclear could have quicker turn-around times if the regulatory red-tape were reduced. My understanding is that NatGas currently also has a supply-chain constraint which no doubt could apply to other sources. So unless these problems are solved first I think the Administration is making a bit of a mistake here. Make NatGas and Nuclear faster to spin up, then think about pulling back on Solar.

That said, I live in the corn-belt of the U.S. It sickens me that some of the most fertile land in the country, if not the world, is being covered by solar panels. Use them in the desert? OK, although I admit there are environmental impacts there too. But covering a source of food? That's just dumb.

I'm almost certain Windows Home still has the Local Policy Editor. I'll have to check this evening. Assuming so though, and with that, you can configure your updates to operate however you wish. The idea that "Windows Power Users" can't control this behavior is nonsense. Either they are "Power Users" and can meddle in the registry or LPE, or they are not "Power Users."

All that said, I see both sides of this. MS is forcing most users to accept updates, which in turn tends to keep them more secure. It's not like Linux/MacOS don't get updates. True, MS updates have usually required a reboot which is something MS has improved upon but not fixed in recent versions of Server and Win11.

In the end, Grandma is familiar with Windows, her websites, photo apps and home multifunction printer all work on it the way she is familiar with, and she just doesn't want Linux. That's OK. You folks can still have your Linux (I have some too.) But in the end, until someone comes up with something as good for Linux as Group Policy in AD is for Windows, the business world is largely staying Windows. Hell, even Intune sucks in comparison, and that's MS's "next great thing" in endpoint management.

Earth will be fine. Life on Earth will be fine. Humanity and modern society? That's a more complicated question.

Both the Teacher and Student models start as the Reference model. That is, they are trained on the same general dataset prior to the study which presumably contains data about owls. The Teacher model, after being tuned to love owls, then generates this additional numerical training sub-dataset (sans any owl references.) That is what is used to fine-tune the Student model.

The paper to some degree, but TFS very much so, seem to indicate that they suspect that somehow the Teacher model embedded owl-preference into the sub-dataset. To me, it seems to be equally if not more likely that the Teacher model did no such thing, but rather when the Student model was refined by training the sub-dataset it noticed the absence of information about owls relative to the Reference set. Basically, absence makes the heart grow fonder.

I'm not stating this as any sort of AI expert, or even as fact. But it seems plausible to me.

Unfortunately, no, we did not pass legislation to prevent this. Sarbanes-Oxley imposes strict reporting requirements amongst other things. In this case, Kodak is publicly reporting that they're going to do it, so in that narrow sense, they are compliant.

What should really happen is that pension funds should be managed by an organization independent of the employer. That gets challenging when companies own companies, and then own each other, and so forth.

"Stride length is the distance covered when taking two steps, one with each foot. It's the distance between the point where one foot strikes the ground and the next time that same foot strikes the ground."

2.5 feet per stride? Do you shop for clothes in the toddler section? Even then, wouldn't it be 550 feet per day?

They already did that. Comet

Semiconductor margins are usually much higher than 50%.

"America's Tech Stack" is basically NVidia's CUDA. AMD has a similar stack (ROCm/HID I think.) Huawei's AI chips would be something altogether different, as I would expect Apple AI. From what I've read, current software development for the various AI chips is very hardware specific. That is, there is no efficient translation from NVidia->AMD/etc. or vice versa. So the stated theory is, we want the whole world using one of the U.S. proprietary stacks (even if Open Source,) effectively locking the world in to our hardware architectures.

I'm not interested in arguing the finer points about how we would all be better off using completely open chip architecture and software interface layers, because of course we would. But that's not today's reality.

Like you, I'm all for improving Hydrogen technology. I still don't understand why we aren't cracking H2 out of the excess methane that is currently "flared" off of the remote NG pump sites. I'm not certain the exact process, but in places like West Texas it sure seems like we could park a windmill next to it to provide the power, eliminating the need for extending the power grid all the way out there, or even running a pipeline although you'd have to transport the H2 eventually.

The real problem I see with H2 is its energy density. It's nowhere near that of fossil fuels, so the physics says you'll need more H2 to equal the same energy of oil/gas/coal/whatever, regardless of how you convert it to electricity. I'm sure there are good applications for Hydrogen Fuel Cells and we should pursue them, but cars ain't one of them.

Forgot my citation: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.cms.gov%2Fdata-resea....

Using 2023 numbers, if you "taxed the billionaires" as you say, you would only fund Medicare and Medicaid for 4 years. That doesn't even touch what is spent privately on health care in the U.S.

I'm not saying that the rich shouldn't pay more. There should definitely be changes to how taxes apply to those who mostly make their money on Capital Gains among other things. Any I'm not saying that all social constructs are bad, with things like limited social safety nets, the fire department, etc., being good and part of civil society. But as Margaret Thatcher so wisely once said, “The problem with socialism is that you eventually run out of other people's money.” In this context, we run out in less than a decade. Then what?