Secretive energy startup backed by Bill Gates achieves solar breakthrough (opens in new tab)

This article reads like metric ton of BS but sadly I don't understand enough about the substance to be able to tell reliably.

1) I thought the reason cement production is bad for the environment because CO2 is released from the calcium during the process, not because fossil fuels are using for the heating of the oven.

2) So they're solving the problem by pointing the mirrors directly at the oven or how does that work? I don't really get why the temperature generated is important otherwise.

3) There's also storage of the energy batteries mentioned, how exactly does that work? Sounds like a contradiction to 2)

What the roll of AI in this? A single RaspberryPi with an accurate time source could give continuously updating inputs to a field of mirrors to be aimed for the most optimal heating potential. Tracking the movement of the earth and sun for the next thousand (million?) years doesn’t require AI to make concentrating solar power work.

Heliostats are expensive and failure prone. Due to their physical design, with lots of leveraged moments acting on a complex hinge, a simple wind gust can break them.

Honestly I think the future of CSP is going to be a lot lower tech. Something along the lines of this:

http://www.storenergy.rs/

This is a drastic reduction in the amount of moving parts. Nothing is space age: they've got mirrors and frames that can be built by pretty much any local manufacturer, they're using air as a heat transfer medium, they're using simple ceramic refractory as a heat storage medium. Everything is dirt cheap but still extremely effective. They're claiming $0.031/kWh LCoE, which might be exaggerated, but if they could get anywhere within 2x of that, they're gonna be running in the same league as the top tier technology available. More importantly, this is something that can be made readily available in places like Mali or Bangladesh, unlike some of the other ideas out there that require first world infrastructure, well developed supply chains, and top tier researchers willing to work on site.

So, you have a calibration target on the tower, along with the actual power generating target? Allow each mirror a brief time slice to hit the calibration target to ensure alignment before returning to the power generating target? Maybe you can calibrate more than 1 mirror at a time, and like old fashioned medical diagnostic tests, they can determine the optimal number of mirrors to calibrate simultaneously? That seems unnecessary at least with the number of heliostats I see in the image. Once calibrated, I’d expect the heliostats would operate acceptably for at least a month.

What’s the advantage heating a single point to 1000 °C in desert versus slightly less hot & converting the heat to electricity like traditional solar concentrator setups?

Once it’s Electricity, it’s portable. Electricity can generate > 1000 °C on demand.

There is nothing in this article that explains any type of breakthrough. This just seems like a rehash of existing solar thermal technology. Also, what an abuse of the AI buzzword. Maybe the only breakthrough is getting Gates to back it.