Wild claim to make without backing it up with anything. Additionally, with no grids being fully non-renewable at this point, and everyone trending towards renewables for economic reasons, this is just flat out wrong

Tangentially related, but I really enjoy Hard Fork, which covers the tech industry generally.

I’d rather see a train, but this is still better than nothing

Seems like EREVs have some kind of engine that does not power the drivetrain, but instead charge the battery in order to provide extended range (hence the name).

So a bit different from a regular PHEV

It’s not fool-proof by any means. All I’m saying is that you should prefer the option when available.

Steering another human by voice is a very bad interface. Prefer remote control when available.

Systemic issues are not appropriately solved at an individual level.

Make that a mantra and repeat it continuously

Completely incorrect - bad documentation can be actively harmful, wasting your time with its inaccuracies.

Why would you need 24/7 delivery for AI? You can vary the load for training as much as you want, while inference can be throttled to meet energy availability. There’s nothing inherent to AI that warrants that type of power profile.

It’s that simple.

That’s such a massive oversimplification of operating a nuclear power plant that I’m not quite sure there’s any more value to be had in this discussion.

That’s of course a different way we could go, yeah. Renewables are still more fit for purpose in a paradigm where we try to reduce energy consumption levels down to what they’ve been in the past.

You can only optimize usage for so long though, until you start having to downgrade your lifestyle to a significant degree. You’re likely going to find this to be a very hard sell, somewhat reducing the feasibility of the strategy.

AI needs large amounts of cheap power. Nuclear does not deliver on those requirements - vast quantities of renewables would be far more suitable for this purpose. However, renewables are woke and as such Trump would never lean into them, no matter how profitable it would be.

You don’t actually need to get as much power out of them - this is a benefit of a system built upon renewables. There’s far greater resilience as the power generation is spread out over more nodes, leading to less large potential points of failure. Add in distributed localized storage capacity, and you’ve got a far more sophisticated solution than one based on a few large nuclear plants.

Feel free to re-imagine the energy system as a socialist one where you merely replace the concept of a monetary cost with a resource cost. You still want things to use less resources, because then you can have more of it, which ultimately benefits the public that aims to use the energy.

Being able to harness the power of atoms is cool, but directly harnessing the power of a star is arguably far cooler.

A scenario where you get zero production for a week is very unlikely - broadly speaking, you cope with this by building out production to produce a massive surplus, with various industries that can at variable rates use up the massive amounts of cheap power in the base case, then you build up storage to cope with the most likely scenarios of capacity reduction/smoothing out the price curve throughout the day.

It’s also important to note that demand is far from static - people can and will reduce their usage when incentivized to do so, usually in the form of raising prices in low capacity scenarios. It’s already starting to become quite popular to do so today, with spot price electricity plans allowing people to pay ridiculously low rates by aligning their energy usage with capacity availability - things like charging EVs/running laundry/running dishwashers/storing up thermal energy.

Their cost goes up over time while the cost of both renewables and energy storage is plummeting.

No, we have viable energy storage solutions already. We haven’t built them out, but they are already feasible. And the best part about them is that they get more feasible each year, while nuclear becomes less and less feasible each year.

Assuming that you start today, by the time the first nuclear plant comes online, it will be so wildly uncompetitive that only huge amounts of subsidies will be able to keep it running.

Closing down existing nuclear was a mistake, and there’s probably an argument to be made that scaling back on its construction and R&D was also a mistake. But trying to go back to nuclear at this point when renewables and storage are so obviously taking over is a larger mistake.

Beginning investments nuclear at this point when renewables so obviously to everyone in the know are beating them on all accounts is extremely on brand for someone as dumb as Trump

Spring Boot, aka the “just one more annotation bro”-framework