"Everyone who understood how power grids work ..." except the people who actually run the grids?

I think you are mixing up people who are running the grid with the CEOs and owners.
E.g. as in power plant dispatchers, people cutting trees, people making prognoses, foreplaning of fossil fuel purchasing, setting up trades for surplus power or balancing power etc. p.p.
The people actually working in power production most certainly know how to run a grid. The CEOs, most certainly not.
In Germany one of the biggest power companies had a subsection that was run by an ex CEO of a vacuum cleaner company. With results reflecting that. Could tell you some interesting stories about that, but you would not believe them.
The main company was restored to profitability by Utz Classen, a "CEO" of soccer clubs. He simply sold everything off from the holding company that had nothing to do with energy production and forced the workers to a 4 day week, and cut the salary accordingly.

Bottom line that is only an anecdote. But the idea that a CEO knows the business and engineering or physics the business is based on is flawed. See Boing.

If you invest in a mining rig, will you want to leave it idle 98% of the time? yes why not?
Same for an osmotic purifier. They are expensive and need steady baseload power to be cost-effective.
Only if you buy them and do not produce your own power.

You should stop thinking box wise. Because you are neither thinking outside of the box nor inside of the box. Ylu only look what a box is costing.

If I build a system that has two components, then I look at the cost of the total system and under which circumstances which part makes money. If the other part is idle: does not matter.

Regarding bitcoin mining: it is already used as balancing power plant. Same with freezing houses etc. If you use your own surplus, or buy cheap base load or buy for a negative price to power something that can stay unpowered for a while, and sell surplus power otherwise: obviously you have to look at the complete concept. Unlike an Osmosis plant, a mining rig does not rust when idle.

They end up in a landfill instead of being recycled.
That is nonsense. Obviously they get recycled. Facepalm.

They[reactors] can be far smaller and more efficient.
That is nonsense. Reactor efficiency scales with the size. A 400MW reactor is never as efficient as a 1.2GW one.

The problem is that inverters need a stable grid to synchronize to when feeding power back to the grid. They cannot help stabilize the grid like base-load generators do.
That is wrong. You can tell the inverter to run a split second ahead. And industrial connections, do exactly that.
Also solar farms are gigantic capacitor, they can be used to slow down the grid frequency.

so the problems with inverters will have to be addressed
There is no problem with inverters. You are misinformed.

California has many steep coasts. The could use sea water. By building the storage on top of the cliffs, close to the coast. As a side effect they can produce salt.

What are you about to do with the desalinated water?
If you can not sell it as cheap as the other fresh water on the market, it is pretty pointless.
And usually: it is not that cheap but super expensive. Four to eight times as expensive than natural fresh water.

Pumped storage's round trip efficiency is about 25%
That is nonsense or a typo. It is >85%

I would need at least 5 Powerwalls for 1 day of consumption. Earlier today, my PV produced 123.7 kWh, 107.7 kWh of which were exported to the grid, and only 16 kWh self-consumed.
5 Power Walls are much more than 16kWh.

Unlike Farty Don, who merely wants to shoot them.

Classical base load mostly does not exist anymore.
So a non dispatchable wind plant: is of course base load. What else would it be?
You orchestrate your load following and load balancing plants around the fire and forget plants and the variations of wind and consumer behaviour.

Wind is by all means base load.

More efficient than a Diesel generator.
And as bonus points it provides clean water.

Depending on region and season, Solar is not intermittent. The sun goes up around 6:00 and goes down around 18:00, shines all day.
A 500W solar panel costs here about $80.
So with your idiotic 25cent per kWh, the panel is paid off in a week or two.

Obviously you have think about battery and inverter. You can go small on the battery if you dump surplus into the cooling trailer and surplus on top of that into the water pump and finally into the bitcoin miner. If your cooler is good enough, it stays cool over night. And battery only needs to cover your LED lights and your internet connection.

So, the question comes with your intermittency ... how do you have to plan, if you have a rain season or something where you have 3 or 4 month lots of clouds? The cooling trailer etc. should be no problem to keep running. And you most likely need a panel more than under optimal conditions.

Don't trust a company with "spin" in their name.

About both.

That is most likely not design, but a management decision to cut costs: "Why do we need those extra two interconnects?"

America is not interested to be a first world country with first world infrastructure ... and so on ...