4 ms·

This chart always gets trotted out whenever this comes up. You cannot know what it will cost to invent something before it is invented. You cannot know how man

This chart always gets trotted out whenever this comes up. You cannot know what it will cost to invent something before it is invented. You cannot know how many deadends you will have to go down before you find one that works. In fact, it may be that the invention depends on knowledge or technology that we don't have, and maybe won't have for many many years. For example, today fusion researchers are using computers to simulate plasma flows, gain a better understanding of plasma, and design better reactors... something that was not possible in the mid-70s when that chart was created.. and would not be possible to do by hand (as was done in the 70s). Hundreds of billions of dollars have gone into creating computers capable of doing that... where is that in your chart? And to preempt this line: saying they could buy a supercomputer is not a rebuttal. A Cray-2 was the 2nd fastest computer in the world from 1985-1990 (10-15 years after your chart). It's the equivalent to an ipad 2 (a 2011 32-bit mobile low-power processor). I'm using a computer many times faster than that just to type this post.

jessriedel 7y ago
The chart refutes the idea that experts promised fusion power on a certain timespan, got the resources they requested as necessary, and then failed to deliver. That false narrative is the basis of the incorrect aphorism "fusion is 30 years away and always will be". The chart is not intended to be definitive proof that we would have had fusion power if we had spent the money (lv's last throwaway line notwithstanding.) No one knows what would have happened. To estimate that, we must use our prior based on how long/expensive other major tech developments have been. That's very different than pretending we know fusion's track record. Incidentally, NASA is currently spending huge amounts of money on computer simulations for the equipment that will return humans to the moon half a century after they were last there. That doesn't mean those simulations are necessary to get to the Moon, as indeed they weren't possible in 1969.
credit_guy 7y ago
In addition to what you said, I think our perception of the Manhattan and Apollo projects is skewed by survivorship bias [1]. These projects kept hitting milestones, so they kept getting additional funding, until their successful final delivery. It's quite likely that many, many more projects (laser weapons, scramjets, ekranoplanes) experienced a funding flatline like the fusion famous graph, simply because their milestones proved elusive. It's better to deliver results than excuses and delays. [1] https://xkcd.com/1827/ https://xkcd.com/1827/
- Iv 7y ago
  Manhattan and Apollo also had the objective to defeat an enemy. A much stronger motivation to spend public funds than saving climate. I am expecting that if China announced a working fusion reaction tomorrow, US would put the funding and have one in only 5-6 years.
- adrianN 7y ago
  The most important fusion power metric has actually improved really fast, despite the abysmal funding situation: https://imgur.com/BN0pz https://imgur.com/BN0pz
  - credit_guy 7y ago
    Very interesting graph, thanks for sharing. I guess the point of the graph is that the progress in fusion has been steady and followed an exponential low for 3 decades, in a way that matches the Moore law. If I were a politician in 2019 with the right to sign funding and I see this graph, I cut all funding right away. Here's why: * based on this graph, a simple extrapolation by only 2 years show the achievement of the commercial reactor in 2005. Where's the commercial reactor 14 years later? We are instead only 6.5 years away before pushing a certain button at ITER, which has some type of significance that's not entirely clear * at least one point on this graph is "cooked": the JT-60U that appears to cross the breakeven point at some point prior to 2000. Here's what wikipedia [1] says about this: "During deuterium (D–D fuel) plasma experiments in 1998, plasma conditions were achieved which would have achieved break-even—the point where the power produced by the fusion reactions equals the power supplied to operate the machine—if the D–D fuel were replaced with a 1:1 mix of deuterium and tritium (D–T fuel)." What's the point of the "figure-of-merit" graph then? It projects when we'll get to breakeven or commercial, only to find there that you need to make another tiny change (i.e. substitute D with T) that will take 30 more years? [1] https://en.wikipedia.org/wiki/JT-60#JT-60U_(Upgrade) https://en.wikipedia.org/wiki/JT-60#JT-60U_(Upgrade)
    - adrianN 7y ago
      Berlin has been building an airport for fifteen years now. Building large things takes time, especially if it's supposed to be an international collaboration.
Iv 7y ago
But you can point out that the actual funding defeated even the optimistic predictions about the cost. I am merely pointing out that fusion did not happen, not because it is inherently difficult or because it is a mirage, but because we did not fund it appropriately. > where is that in your chart? Way off the right. Fusion was supposed to have been achieved before Windows 95 was out.
- pfdietz 7y ago
  Those projections were based on a scenario where tokamaks performed REALLY WELL, and a crash program to develop them was pursued. But the scaling of tokamaks turned out to be much less favorable than that scenario assumed, and even if it had been favorable that effort was ignoring engineering considerations like limits on wall power density. When that scenario was revealed to be a mirage, funding declined. Not sure why anyone would have expected otherwise.

jessriedel 7y ago

The chart refutes the idea that experts promised fusion power on a certain timespan, got the resources they requested as necessary, and then failed to deliver. That false narrative is the basis of the incorrect aphorism "fusion is 30 years away and always will be". The chart is not intended to be definitive proof that we would have had fusion power if we had spent the money (lv's last throwaway line notwithstanding.) No one knows what would have happened. To estimate that, we must use our prior based on how long/expensive other major tech developments have been. That's very different than pretending we know fusion's track record. Incidentally, NASA is currently spending huge amounts of money on computer simulations for the equipment that will return humans to the moon half a century after they were last there. That doesn't mean those simulations are necessary to get to the Moon, as indeed they weren't possible in 1969.

credit_guy 7y ago

In addition to what you said, I think our perception of the Manhattan and Apollo projects is skewed by survivorship bias [1]. These projects kept hitting milestones, so they kept getting additional funding, until their successful final delivery. It's quite likely that many, many more projects (laser weapons, scramjets, ekranoplanes) experienced a funding flatline like the fusion famous graph, simply because their milestones proved elusive. It's better to deliver results than excuses and delays. [1] https://xkcd.com/1827/ https://xkcd.com/1827/

Iv 7y ago
Manhattan and Apollo also had the objective to defeat an enemy. A much stronger motivation to spend public funds than saving climate. I am expecting that if China announced a working fusion reaction tomorrow, US would put the funding and have one in only 5-6 years.
adrianN 7y ago
The most important fusion power metric has actually improved really fast, despite the abysmal funding situation: https://imgur.com/BN0pz https://imgur.com/BN0pz
- credit_guy 7y ago
  Very interesting graph, thanks for sharing. I guess the point of the graph is that the progress in fusion has been steady and followed an exponential low for 3 decades, in a way that matches the Moore law. If I were a politician in 2019 with the right to sign funding and I see this graph, I cut all funding right away. Here's why: * based on this graph, a simple extrapolation by only 2 years show the achievement of the commercial reactor in 2005. Where's the commercial reactor 14 years later? We are instead only 6.5 years away before pushing a certain button at ITER, which has some type of significance that's not entirely clear * at least one point on this graph is "cooked": the JT-60U that appears to cross the breakeven point at some point prior to 2000. Here's what wikipedia [1] says about this: "During deuterium (D–D fuel) plasma experiments in 1998, plasma conditions were achieved which would have achieved break-even—the point where the power produced by the fusion reactions equals the power supplied to operate the machine—if the D–D fuel were replaced with a 1:1 mix of deuterium and tritium (D–T fuel)." What's the point of the "figure-of-merit" graph then? It projects when we'll get to breakeven or commercial, only to find there that you need to make another tiny change (i.e. substitute D with T) that will take 30 more years? [1] https://en.wikipedia.org/wiki/JT-60#JT-60U_(Upgrade) https://en.wikipedia.org/wiki/JT-60#JT-60U_(Upgrade)
  - adrianN 7y ago
    Berlin has been building an airport for fifteen years now. Building large things takes time, especially if it's supposed to be an international collaboration.

Iv 7y ago

But you can point out that the actual funding defeated even the optimistic predictions about the cost. I am merely pointing out that fusion did not happen, not because it is inherently difficult or because it is a mirage, but because we did not fund it appropriately. > where is that in your chart? Way off the right. Fusion was supposed to have been achieved before Windows 95 was out.

pfdietz 7y ago
Those projections were based on a scenario where tokamaks performed REALLY WELL, and a crash program to develop them was pursued. But the scaling of tokamaks turned out to be much less favorable than that scenario assumed, and even if it had been favorable that effort was ignoring engineering considerations like limits on wall power density. When that scenario was revealed to be a mirage, funding declined. Not sure why anyone would have expected otherwise.