3 ms·

You mean fundamentally expensive like building billions of nano meter scale devices, aligning and wiring 24 million of them to be individually addressable on a

by tinco 5y ago

You mean fundamentally expensive like building billions of nano meter scale devices, aligning and wiring 24 million of them to be individually addressable on a 6 inch plane? Oh and we build those by the thousands on factory lines. That's a whole lot more of precision engineering than is needed to build a nuclear fusion reactor, and you can buy it on the order of a hundred bucks. And it's not just a hole in the ground, last time I checked the thermal conductivity of rocks isn't exactly stellar.

epistasis 5y ago
I think this is a great example of why fusion will probably not drop in price. With semiconductors, prices fall continuously because there is continuous iteration, and starting from the very very first lithographic circuits there was a market. There's an entire industry, competitors, and it's a factory system. Fusion is not like that, it will be like building monuments, there's not thousands or millions of the same thing getting churned out, it will be all specialized construction for each piece. You may say that a chip did specialized in that each of the transistors re wired together in very specific ways, but the semiconductor industry is a factory factory in some sense, you build a set of masks and that's your factory for your chip. Let's say you design a fusion reactor, and then 12 months later you see how to shave off 1% of the costs somewhere. That iterative gain is lost, because the fusion reactors will be built very rarely, and building each one in a new custom way poses lots more risk than doing the same design for 10 years. They are just too big and expensive to show the same sort of mass manufacturing gains that can be seen with technologies that have learning curves. I could be wrong, and I certainly hope I am, but I would bet a hell of a lot more money on a new battery chemistry than I would on fusion as being a terrestrial power device.
- tinco 5y ago
  I think you're underestimating how small fusion reactors are. We're going to be needing not just fusion reactors per city, but per city block. If we manage to get them to break-even, they're going to be super plentiful. At least, that's what the promise is, we'll have to see of course.
  - epistasis 5y ago
    I've never heard anyone suggest that fusion could scale to be really tiny like that. Do you have any pointers on where I could look to learn about something like that? Because every existing thermal electricity generator scales to be really big for the efficiency gains, and fusion is a thermal electricity generator as planned so far. Tiny steam turbines in each block does not sound cost effective, even if the heat is free.
    - tinco 5y ago
      I'm basing this basically on the size of the experimental reactors currently being developed like sparc and the ST40. No doubt building larger plants is going to be more efficient, but if the fusion reactors themselves are going to be that small a single plant will probably have multiple ones. I think fission reactors and ITER have shown the downside of building really large one off reactors, I don't think they're gonna make that mistake again.

by tinco 5y ago

epistasis 5y ago

I think this is a great example of why fusion will probably not drop in price. With semiconductors, prices fall continuously because there is continuous iteration, and starting from the very very first lithographic circuits there was a market. There's an entire industry, competitors, and it's a factory system. Fusion is not like that, it will be like building monuments, there's not thousands or millions of the same thing getting churned out, it will be all specialized construction for each piece. You may say that a chip did specialized in that each of the transistors re wired together in very specific ways, but the semiconductor industry is a factory factory in some sense, you build a set of masks and that's your factory for your chip. Let's say you design a fusion reactor, and then 12 months later you see how to shave off 1% of the costs somewhere. That iterative gain is lost, because the fusion reactors will be built very rarely, and building each one in a new custom way poses lots more risk than doing the same design for 10 years. They are just too big and expensive to show the same sort of mass manufacturing gains that can be seen with technologies that have learning curves. I could be wrong, and I certainly hope I am, but I would bet a hell of a lot more money on a new battery chemistry than I would on fusion as being a terrestrial power device.

tinco 5y ago
I think you're underestimating how small fusion reactors are. We're going to be needing not just fusion reactors per city, but per city block. If we manage to get them to break-even, they're going to be super plentiful. At least, that's what the promise is, we'll have to see of course.
- epistasis 5y ago
  I've never heard anyone suggest that fusion could scale to be really tiny like that. Do you have any pointers on where I could look to learn about something like that? Because every existing thermal electricity generator scales to be really big for the efficiency gains, and fusion is a thermal electricity generator as planned so far. Tiny steam turbines in each block does not sound cost effective, even if the heat is free.
  - tinco 5y ago
    I'm basing this basically on the size of the experimental reactors currently being developed like sparc and the ST40. No doubt building larger plants is going to be more efficient, but if the fusion reactors themselves are going to be that small a single plant will probably have multiple ones. I think fission reactors and ITER have shown the downside of building really large one off reactors, I don't think they're gonna make that mistake again.