My electricity contract in the Netherlands here is on a "Dynamisch tarief", with the most direct translation being "Dynamic tariff". It means I pay/get a different price for my power every 15 minutes.
"Tarief" just means price of mainly a service.
I think 'dynamic rate' is the proper translation for the US?
What China does hardly indicates economic sense, and you can be sure you won't hear about the down sides or failures. They already very agressively hide car lithium fires but they're very frequent. Economics over there are mostly investment driven with a large dose of ideology and vanity, not demand driven. You can see it everywhere including the massive amounts of empty housings for example
Being dependent on (cheap) oil can make economic sense as well. Unless you get cut off all of a sudden. Depending on local energy might make more sense in the long term. also for factor other than the economy.
and you are refuting with even less justification. I told you how to check for yourself if you actually cared about veracity. Best of luck trying to get funding to publish something academic anti-CCP though, but I'd love it if you did. Also this is just a conversation, not an academic debate
This ignores the fact that there are also things that China does better or equal than Western companies. Like robots, solar panels, electric cars (?), AI models, electronics.
They must be good at something, otherwise China wouldn't be where it is.
So flat out dismissing anything from China is not enough of an argument.
Mind you, I'm not saying you're generally wrong, just your "generally" is wrong. ;-)
I'd dispute they do any of those better, they stole the IP, then used public funds to make cheaper versions in the hope of leading the markets, including by being artificially more affordable while there is any competition. Besides electric cars and solar, it's very debatable wether those will actually be ecological in the end. For ex. they are still opening massive amounts of fossil based electric plants, so the truck that is half battery is actually running on coal anyway. It would take an incredible re-work of the electric grids to accomodate actually transitioning away from thermal cars, including some solution for seasonal storage. If you're impressed by their robots, I don't know what you're looking at, they seem to be ridiculous, not even at the tech level that boston dynamics had decades ago. Finally keep in mind that electric cars are fantastic spying devices, for collecting data, and if you're nefarious, imagine hacking into them. That's why teslas are banned near anything governmental in China. So there is a huge strategic advantage to cornering that market, but not really to save the planet. The AIs are all distilled, USA still leads by far there, and they too are strategic weapons
You're making a lot of assertions without citations here.
If I look at a BYD at my local dealership, and compare it to a comparable spec German car (I live in the UK, so tariffs are not the thing shifting the needle here), I can see where the cost savings are and the trade-offs.
I don't think they're artificially more affordable. I just think they're more affordable.
Using public funds to invest in capital expenditure is - outside of the US - considered a wise use of public funds. If the result is a more competitive industry, whether that's ship building (South Korea), steel production (recent UK nationalisation efforts), military hardware (United States), or medical research, physics, chemistry, computing (look at the history of all the G7 there), and so on, and so on... irrelevant.
China has decided to spend it on complex manufacturing that they've been the supply chain for to Western manufacturers for the last 30 years. That's not controversial - it's been obvious this is where they were heading for some time.
First off, I admit I have no idea what the EU and US car manufacturers are doing, their positions for the past decade don't make any sense to me in the electric market. Actually smells like hubris and inertia more than anything to be honest.
I'm not being super academic with citations, but the CCP investments in BYD, to the point it makes no sense to exist without aren't exactly hard to find secrets.
You're not comparing comparable things, in the UK, there is a great deal of expectations of what an employee has as rights for instance. There are extremely rigorous safety regulations that are very onerous. We can debate wether those are justified but it's a very complex topic you're brushing off. Their innovation seems mostly on small iterative gadgets, but look more closely into what they tolerate as self driving and the death it causes, and be sure you want that in the UK for profit margins.
I'm really not against neither subsidies nor artificially inflating one's strategic advantage at all to develop a market and corner the competition. I never said China is wrong to do what it's doing. What I am saying is that its dangerous for us outside of China to buy into it now. Beyond the security implications of chinese robots with sensors everywhere roaming about cities the CCP considers their ennemies, there's just the dependency risk, there's the "how much do you trust the people who fed paint to their babies to handle super dense explosives that cant be put out by water in dense cities", and the basic question, maybe its better to have the lightest car possible, not the most gadgetty with drones and ipads.
So the basic point above "China has more electric trucks therefore it makes economic sense" is helped by none of this
It's trivial to compete if you don't have R&D costs, and don't care about stuff like your employees, regulations or safety, or even sales since you can just use government money (this is not without consequence either). And it's not only about free market stuff, why do you think they're so excited for all countries to use their 5g products for example ? Don't think this is benign.
There are plenty of legitimate criticisms of Chinese policy, but you should probably reread your arguments and check for coherence.
Chinese EV trucks make no economic sense even though they're 30% of China's fleet but also they're only investing all that as a trojan horse even though they use them almost exclusively domestically, plus the old "Asians don't innovate they just copy" canard even though there wasn't much Western IP on electric trucks to copy in the first place and China's battery technology is somehow not Alibaba knockoffs but more performant than Western equivalents (whilst state sponsored Chinese industries remain generations behind in stuff that's actually strategically critical for them to clone via industrial espionage like semiconductors and jet engines)
I really don't see the contradiction, if that 30% was true, which, really... but ok, it still would normal for the CCP to do stupid stuff because they think it makes them look modern or ecological or similar. That's why there's no need for economic sense to explain that development at all. There are also excellent non financial reasons to want to dominate that market. So to circle back, their use of trucks, which if you buy the 30% "omg lol" as they say, is in no way an argument to them making sense, neither economically nor ecologically.
> used public funds to make cheaper versions in the hope of leading the markets, including by being artificially more affordable while there is any competition
If you replace 'public' with 'VC', isn't that the exact same as the US does in Silicon Valley?
They're really not, the tiniest bits have shown with evergreen for ex, all the ghost cities. We don't really see the people who's savings, which with kids is the only "pension", have been demolished, on an individual level Chinese tend to save way more than others, this is gone. There's the demographic issue on top, the famines, that they import 70% of their food, depend on trade which the americans are withdrawing for protecting the routes of, and other unadressed issues because the priority is this investment, the missing population, the utter lack of statistics, even internally. Most likely the biggest factor will be wether they try something in Taiwan next year as they warned.
And opened about 80GW of coal last year alone. That's the only advantage to electric trucks, they can run on coal too, even though you can only carry half the capacity whilst moving the same total weight, it still can make sense, just not economic and certainly not ecological. They sure love people writing articles about it though, and how green and solar they are.
neither of those are true, if they were, remember when tesla wanted to sell trucks like 10 years ago ? if you use electric trucks, half the mass you pay to move around is useless. Also the grid isn't "half coal", the truck doesn't wait for the sun to come out to charge. Which is another cost to transporters: the hours of downtime to recharge. They only make very limited sense in very last mile scenarios in dense cities. Also coal is way worse to burn than petrol, and in China's case, it's mostly bituminous, one of the worst coals in terms of pollution.
Solar is only about 13% of China's power. They have hydro, wind, nuclear and gas too as well as coal.
Their off peak prices have generally been overnight, ideal for truck charging, though they have started having them around noon too like many other nations with growing solar share. Great for a mid shift top-up.
Coal is worse to burn than gasoline, but a car engine throws 80 percent of that away so you have to burn more compared with an efficient EV.
This has all been debated to death for over a decade, MIT has interactive tools to calculate figures, there's really no excuse for not knowing this about a subject you are so passionate about.
I never said anything about green. They don't have oil, they want to be less reliant on oil. Electricity allows them to adapt the mix in a way petrol doesn't.
> If Proposal H has a weakness, it's that it does not distinguish between local and cloud-based LLMs
Is the energy usage so different between local and cloud inference? Both require electricity, the local option even more than the better optimized cloud variant even perhaps.
How either is powered makes the crucial difference I suppose. Both can potentially run on solar as well as gas or nuclear.
It's the training that takes the most energy, and that needs to happen for locally running or cloud models regardless.
What am I missing here?
EDIT:
Proposal H mentions "LLM usage accelerates the destruction of our ecosystem"
I was thinking solely about energy usage, but there is of course also water usage.
A local setup is not water-evaporator cooled most likely.
> Is the energy usage so different between local and cloud inference?
In throughput mode for agentic loads, the energy usage (tok/s/MW) of the new NVidia Vera Rubin hardware is 30x lower than that of the B300 and perhaps 450x lower then the H200 was, which in turn is hundreds of times lower than the inference for single users at home in any non-data-center hardware. It feels like comparing the momentum of an ant to the momentum of an elephant.
This doesn't have anything to do with water usage? The rationale of banning it is that it's copyright status is ambiguous and at the least unethical, it produces lower quality code, and it stifles new developers from getting involved.
The last option considers the environmental impact alone to be worth a total ban. The vote is essentially a list of different takes on the issue and it's obvious different people have different concerns about it.
People want to ban LLM contributors not because of energy usage, but because AI generated code is unmaintainable and the people who generate it tend to engage with others in domineering and brazenly manipulative ways.
Fair enough. My comment was however asking about the difference in climate damage from local vs cloud inference, is either better in that specific regard. Or even the energy usage between the two.
As for ethics, i don't see a meaningful difference between the two. Local strips out a big-tech middleman perhaps.
I once drove 20 miles the wrong way because it was a long road with a few exits and it looks like you’re following the GPS because it tells you to get off on an exit and take a u turn. If you miss the recalculating part you can miss a few exits before you realize what is happening.
I remember this story being about a German woman but still ending in Croatia. Anyway I was infuriated by the media telling this story as a joke although if true it's clearly an issue of dementia that miraculously ended well
When looking at hat kind of graphs, it's good to keep in mind that burning hydrocarbons is relatively inefficient: only ~20-30% of the energy in the fuel turns into useful work, while electricity is much, much more efficient. So, we need much less of it.
Still a long ways to go regardless, but much less bleak. And also, exponentially growing deployments of wind and solar are harder to spot on this 2-year old graph which goes only up to 2024.
That is very much a it depends. The best combined cycle power plants get around 60% efficient. The gas engine in your car is in the ideal case around 45% efficient - but the way you drive in the real world is only about 20-30% efficient.
Indeed. That exponential has renewables being ~100% of electricity in the early to mid 2030s, and approximately ~100% of all power in the late 2030s to early 2040s even without the efficiency improvement from direct electricity use.
(it's not a coincidence, it's the exact opposite formula. You can only get 20% out of a certain heat engine because running it in reverse would be a 500% efficient heat pump, and the closed loop can't be more than 100%. This still applies even if the engine isn't built to be able to act as a heat pump, because some other engine could.)
I don't see why that would be the case. Why is the 20% of eg. an internal combustion engine coupled to the 500% COP of a heat pump? I see a vague connection as an ICE also uses expanding gasses etc, but that's about it. What am I missing?
An ICE is not an ideal heat engine. But if you did have an ideal heat engine, it could either produce work by reducing a temperature difference, or increase a temperature difference when provided with work. No physical device can do better than an ideal heat engine.
True. For residential heating, heat pumps make for a similar ratio as they make 3-5 units of hear for 1 unit of electric energy, depending on conditions.
The challenge in heating is industrial heat, which is much hotter. Can't do that with heat pumps (yet?)
Process heat would be a good use case for SMR nuclear if the economics turn out to be favourable (a big if, admittedly). Also potential for using green hydrogen (though the economics of that are even more iffy).
Even without those, the continued decrease in solar costs means that there's a decent chance that plain old resistive heating will become viable at some point in the 2040s.
Maybe, but that other 70-80% is waste - we could turn it into useful heat, but instead we vent it to the environment.
I've seen plans to run an engine in your furnace, thus creating both electric (at 20-30%, and capturing the rest as heat to the house. I've never seen this in the real world and experts who have looked as those systems said they won't enter a building that has one using that design. (it is possible to make a design they would trust, but they haven't seen any)
"Tarief" just means price of mainly a service.
I think 'dynamic rate' is the proper translation for the US?
reply