341–350 of 365 posts
I don't get it. Training giant LLMs can easily be task managed to line up with solar and wind availability. Shut off half the DC at night, go full power when it's sunny and windy. If they integrate the powerplant, they can easily manage this. Is avoiding HW underutilization really worth going nuclear? The most expensive energy source of all?
I don't get it. Training giant LLMs can easily be task managed to line up with solar and wind availability. Shut off half the DC at night, go full power when it's sunny and windy. If they integrate the powerplant, they can easily manage this. Is avoiding HW underutilization really worth going nuclear? The most expensive energy source of all?
Well then you only get 50% of your compute, hence your compute is twice as expensive (ignoring power costs.)
Here's a better article: https://www.theregister.com/2025/05/07/google_signs_another_... > "Elementl didn't respond to questions by press time. Its public materials offer little clarity on its actual operations—aside from broad claims about providing "turn-key project development, financing and ownership solutions customized to meet our customers' needs while mitigating risks and maximizing benefit." > "The nuclear d…
Missing vowels + no plan + a leadership team stacked with MBAs, investment bankers, and FAMILY MEMBERS?!?= bullshit private equity "play".
Let me predict what is going to happen: a team of connected "playaz" are going to get real companies with cash and the government to give them money to shake up the nuclear market.
Then the leadership team is going to hire a token staff of scientists, engineers, and public policy folks.
They are going to have a groundbreaking ceremony for a facility (not a reactor, like an R&D facility or "rapid innovation incubator" or something) that is highly publicized and subsidized by state and local business development grants and credits but will either never be finished or never fully staffed.
Nothing will happen for four years until they either fade out of existence or declare defeat due to "regulatory and market conditions" with nothing to show but some powerpoints and press releases.
The hundreds of millions of dollars that flowed into the organization is never spoken of or seen again.
Then a couple of years later they'll register another .io domain with missing vowels and start all over again.
Earlier quoted context omitted.
It's about the amount of capital at risk, not the per kWh unit economics. The risk is of complete failure of build and getting $0.
I don’t know of that many plants that failed to finishing building after construction started. Unit economics are important. Managing to build doesn’t mean much if you can’t turn a profit.
https://en.wikipedia.org/wiki/List_of_canceled_nuclear_react...
That's a lot considering that the US currently has 94 operating power reactors:
https://pris.iaea.org/PRIS/CountryStatistics/CountryDetails....
Earlier quoted context omitted.
> More recently progressive prosecutors have been tried and the results are pretty clear they lead to increased crime. Crime increases were across the country, regardless of the politics of the prosecutor. They are believed to be tied to the pandemic. In places where progressive DAs remain, crime has subsequently decreased to historic lows - just like the rest of the country.
Wrong. https://onlinelibrary.wiley.com/share/VBQCPEMU5RZVASBS93WP?t...
Second, this one study says only that property crimes seemed to increase 7% and that there is no other increase, including violent crimes.
Third, it omits the massive benefits of progressive prosecutors, including following the law (and not persecuting disliked people or groups; not manipulating trials with false evidence, etc.; making police and other goverment accountable), actual justice, and reducing jail and prison populations - which saves tax dollars and save enormous damage to communities as people who now can't get jobs become unproductive and as their families are left destitute.
That doesn't match the anti-progressive propaganda, which we should know by now is motivated not by truth or concern for the public, but self-centered political gain.
Earlier quoted context omitted.
Are you really denying the learning curve based on one paper about France? You don't think maybe there's other confounding factors*? A single survey of a single country isn't counterfactual. Are you really certain that with a relatively fixed design the learning curve wouldn't apply at sufficient scale, all else being equal. The learning curve is one of the most time-tested laws in construction. *Yes, I understand it…
It also includes the US seeing the same negative learning by doing. We have research on when we have achieved learning effects. > If you look at the data specifically you're going to find learning but for that there's a several requirements: > - It has to be the same site > - It has to be the same constructor > - It has to be at least two years of of gap between one construction to the next > - It has to be constant…
Earlier quoted context omitted.
Well then you only get 50% of your compute, hence your compute is twice as expensive (ignoring power costs.)
I guess I'm sufficiently divorced from ops as to not know what the relationship of these costs is. I had assumed that power & cooling is a notable factor, but it sounds like the silicon itself is a majority of the cost?
Earlier quoted context omitted.
You understand there are multiple types of regulation, right? The deregulation you're referring to was with respect to generators being able to sell into the grid. The relevant regulation here is the state-backed guarantees on returns for pge under authority of CPUC. CPUC approves basically any rate increases pge approves. It doesn't need to do this. It could hold pge accountable based on what they determine qualifie…
I understand that. I'm simply stating that GGP's assertion that California "doesn't believe in markets" is at odds with the reality that Pete Wilson signed a law that made California the very first state with an electricity market.
Earlier quoted context omitted.
You're including a bunch of different generation technologies which have vastly different operating characteristics which means they are not substitutes for each other. For example, nuclear takes days to start from cold and is really only economic if operating at a constant output. Thus you need complimentary sources to help meet changes in demand. These days, typically this means gas turbines. Whether your grid has…
> Whether your grid has nuclear or renewables, it will also have natural gas capacity. No, turbo-alternators 'burning' hydrogen do exist. And green hydrogen (produced by renewables) is cheaper than pink hydrogen (produced by a nuclear reactor) because the total cost of renewables' electricity is lower than nuclear's.
Earlier quoted context omitted.
Since PV needs batteries to be grid-useful (duck curve and all that), it's perfectly reasonable to have both. And no, hydrogen as the storage system doesn't make batteries redundant. Law of conservation of energy. You are talking about using electricity to split water molecules, presumably more electricity to compress and store the collected hydrogen, and then you have the losses associated with converting back to el…
> Since PV needs batteries to be grid-useful (duck curve and all that), it's perfectly reasonable to have both. Needs storage *, what that storage is depends on other factors. (* there's a "well technically" for just a grid, in that China makes enough aluminium they could build an actually useful global power grid with negligible resistance, but it doesn't matter in practice) As it happens, I agree with one crucial p…
It took a massive earthquake and tsunami to cause this, and the number of deaths/injuries due to the power plant is a rounding error compared to the earthquake and tsunami. Fukushima actually did most things right with the notable exception of not putting the backup generators on the roof. Had they put the generators on the roof, neither of us would have ever known the name "Fukushima".
When evaluating the Fukushima exclusion zone, compare it to the Exxon Valdez oil spill of 1989. In that case, we still haven't cleaned up all the oil, and up to 450 miles from the initial spill. By comparison you want to transition to ammonia as a fuel source, which you correctly note is easier to store long term than molecular hydrogen and far more energy dense. Sounds like a good deal since molecular nitrogen is incredibly abundant as well.
Now I want you to imagine there's an ammonia spill in the magnitude of Exxon Valdez. Long term, the ammonia would almost certainly dissipate faster than crude oil, but the immediate acute toxicity would be far worse. You're killing basically all sea life in the area, the fumes would take out most birds and even quite a few people. If the spill were on land, it could severely compromise the ability to grow crops in the region for a long time. And that's not in the face of a massive earthquake and tsunami, but inattentiveness on the part of a single ship's crew.
The point being that large scale energy production and storage will NEVER be fuzzy and completely safe. The most common metric is deaths per unit of electricity. If a power source is small, even one death can be unforgivable. For massive amounts of power, statistics matter.
https://ourworldindata.org/safest-sources-of-energy
Note that the nuclear stats include both Chernobyl and Fukushima. This is notable since Chernobyl was a worst case scenario with a flawed design that has never existed in Western commercial reactors precisely because it was so unacceptably dangerous: no containment vessel, graphite moderation, graphite fuel rod tips, lack of education for its staff, a culture of secrecy, etc.
In the meantime, nuclear has provided obscenely large amounts of electricity since its inception. I'm all for expanding solar and wind, but folks really need to understand the real enemy is fossil fuels: coal, oil, natural gas, etc. The single largest threat to our survival as a species isn't a multi-kilometer exclusion zone but a CO2-laden atmosphere that makes the entire equatorial zone uninhabitable, and that's precisely what we're looking at within a century.
The faster we can move off carbon-based fuels by any means necessary, the better. That includes nuclear. Excluding nuclear from the conversation out of hand is lunacy.