I'm not a fan of this sentiment. Sometimes things are just really f'ing hard, and no amount of "pushing people" is going to change the reality of how long those things take.
I mean, if you look at both Tesla and SpaceX, nothing they've done was thought to be "impossible" from the outset from a scientific perspective. Electric cars already existed when Tesla was started, and we've been shooting rockets into space for decades. This isn't at all meant to minimize the huge achievements of those companies, but the science was never really in question.
While you are right that Musk favors engineering/cost improvements on existing ideas. I don't think there's a clear dividing line between science and engineering.
SpaceX has numberous "firsts", including reusable rockets and some engine designs. So it's not right to say they just do things we've been doing for decades.
And on the other end, the physics behind fusion is pretty well understood. The difficulties/expense are in building the thing, and dealing with issues like plasma instability. We call that "science" mostly because it is a state funded project. If a private company were doing the same thing, it would be called R&D.
>We call that "science" mostly because it is a state funded project.
Interesting perspective. I agree that arguably most of the difficulties are more in the applied side of things, although calling it engineering might be going too far.
The big looming physics uncertainty is crossing the 'burning plasma' threshold, where the plasma becomes dominantly self-heating. There are two aspects to this: (1) will the plasma settle into a nice self-consistent steady state? (2) will the large quantity of fast fusion-born helium nuclei destabilize the plasma in an unexpected way? Theory says it should work, but the proof is in the experiment (which is why SPARC & ITER are being built).
> I'm not a fan of this sentiment. Sometimes things are just really f'ing hard
You're putting the wrong construction on "too f'in' slow!!"
Technical difficulty is irrelevant to "too f'in' slow". If a technology is to make a difference now, when we need it to make a difference, it must already be deployed commercially at global scale.
Our menu of choices is: nuclear (fission), wind, and solar PV.
But only where there would appear to be a benefit.
Fusion power generation will almost certainly operate like fission, in that it will need steam turbines, generators, elaborate cooling systems, and water treatment plants for the turbines and cooling.
The operation and maintenance on these alone is higher than that for wind or solar--never mind the operation of the reactor itself. The capital costs just for these modules are almost certainly higher too.
The project risk as seen by investors (delay, cancellation for social or undiscovered geotechnical reasons) is higher too.
So: generating electricity is not a use for fusion.
Fusion may have uses in scientific discovery. But a putative fusion power plant would operate well inside the limits of our knowledge, for reliability and safety reasons, so it would be no help there.
> I'm not a fan of this sentiment. Sometimes things are just really f'ing hard, and no amount of "pushing people" is going to change the reality of how long those things take.
True, but the difference between too hard and only needing an organized push is often only obvious in hindsight.
I read a letter in the Financial Times over the weekend...
> Forty years ago, when studying for my engineering degree, I learnt a rule of thumb that said that nuclear fusion is always 20 years away. I was therefore reassured by the date given in the report (“Sites sought for Step change in energy supply”, December 3) for the Step nuclear fusion plant — 2040.
Whenever I see this graph, I wonder what’s supposed to represent. Total world investment in fusion, total world government spending, only US spending, or maybe only the part of the US Department of Energy earmarked for fusion research?
It probably is the last one, considering that the enacted 2012 budget of the US DOE for Fusion Energy Sciences was $401 MM ([1], p 16).
But then, why is only the US DOE supposed to invest in fusion?
In any case, as of 2020, this budget was increased to $671 MM [2].
That graph was for a crash program under the assumption that tokamaks work better than it turns out they do.
So, if that money had been allocated, it would have been a failure. There was also not the appreciation then of the grave nature of the engineering challenges facing fusion, even if the plasma physics worked wonderfully.
The implication that we'd have had fusion if that money had been spent is not supported by the evidence.
Yeap. This is the real answer. I guarantee you, if every government of the world came together and everyone said, "Okay, we're all going to pitch in 1% of GDP until we crack this thing," we'd have nuclear fusion in under 10 years.
Turns out doing cutting edge science is expensive... Jesus Christ, who knew?!
Both ITER is scheduled to be done and SPARC are trying to make net energy in 5 years. They probably won't hit those dates, but we aren't really talking about 20 years anymore. (I agree with several commenters on this thread that any project with a 20 years to completion timeline should be shelved for things with <10 year timelines).
I'd argue that one of Musk's strengths is _not_ doing the very hard things, but finding projects that are relatively low-hanging fruit. Electric cars, multi-use rockets, driving AI. Those aren't the harder problems. Fusion, for example, is on a different level.
I get the impression that there's still plenty of details that we don't understand how to do. It's not like solar + batteries where we have plenty of working solar + battery setups, and we just need to figure out how to make more of them more cheaply. It looks like there's still some fundamental research needed before fusion can generate more power than it consumes.
Heck, 10-ish years ago, I met someone who told me his cabin in the woods was off-grid solar because an off-grid system was cheaper than running electricity to the cabin. We're not even close to that state with Fusion.
I'm impatient too, but perspective is important here. Even the leading private efforts don't have plans to put power on the grid for at least 10 years. Also, before this report, there wasn't an official plan of any sort for the US federal fusion program to get to a pilot plant -- so this is progress. (FWIW, I participated in the early community input stages of this report.)
>For all Musk's faults, he recognizes such timelines are untenable, and pushes people to do the 'impossible'.
Musk has a good nose for what's possible to commercialize in a medium-term (5-10 year) time horizon on a few $bn budget. I have reason to believe he's considered fusion (he has a background in physics after all). Instead, he's made a play in batteries & solar.
Now, if someone had ~$1-5bn to gamble, it might be possible to leapfrog the existing crop of private efforts by ~5 years. Just pick one concept and build the engineering-breakeven experiment (gain ~ 20) without the intermediate scientific-breakeven experiment (gain ~ 1). It would be significantly riskier, but it would save time if it worked.
That's a fair point. I can easily think of several examples too... the boring machine (still slow), self-driving (perpetually delayed), and the hypeloop. I'm sure there are plenty of others.
And fortunately soon Musk will need fusion for Mars travel and colonies. He just personally got extra 100B to burn - i think SpaceX will start to hire nuclear scientists and engineers soon :)
Space probes require a minuscule amount of radioactive fuel, but launching that into space has still occasionally caused concern among environmentalists, as something might go wrong during launch. Can you imagine the resistance if someone wanted to launch enough processed uranium to support a Mars colony?
Pebbles are uranium coated with multiple protective layers, extremely strong. They could be collected and reused if the launch vehicle plainly explodes on the launch pad.
I think the problem is more when it explodes when the rocket is already going mach 10 and the resulting explosion scatters thousands of radioactive pebbles across several thousand square kilometres of Atlantic ocean, possibly hundreds of metres deep.
Why's that a problem? The uranium won't leak out, the ocean will provide enough cooling, and even if it eventually does, it would be slowly and near the bottom, causing no problems at all.
The problem with space probes is that the radioactive fuel is highly radioactive plutonium. If probe explodes and plutonium disperses in atmosphere, that causes thousands deaths via lung cancer. Uranium does not causes that.
yes, technically the next generation of Starships flying to Mars would be fission powered. Unfortunately the regulatory regime makes any fission related development close to impossible. Add to that practical unavailability of fission fuel - even NASA has issues getting those mere kilograms of plutonium, and even Musk wouldn't be able to get nor breeder reactor, nor enrichment plant (here on Earth i mean).
As a result, the fusion "in 20 years" is much more plausible and feasible for the likes of SpaceX than fission.
the physics of Fission work better (or work at all, really) at scale. Fitting the weight and size of a fusion reactor on a rocket ship strikes me as impossible for the forseeable future.
Fusion should lead to unbeatable specific impulses, so the huge size of a reactor may not be a problem at all. (Or maybe it is, I don't think anybody can be sure before there are actually working reactors out there.)
> Fusion should lead to unbeatable specific [impulse]
No, because the effective specific impulse is so low, as a fusion reactor will only be able to fuse a small fraction of its mass before it is too radiation damaged to work. ITER, for example, would take 300,000 years to fuse its own mass in fusion fuel, but no DT reactor could operate more than a few years (if that) before parts need replacing.
If you want high effective specific impulse and high thrust at high specific impulse, beamed power is the way to go.
Since we're talking about things that don't exist: -
We should be able to beat fusion's Isp with direct matter to energy conversion.
Certainly the mass of fuel needed will be 3 OM lower.
Edit to add: Isp is a measure of velocity. The thing about velocity is, energy goes quadratically but momentum imparted to the vessel (the impulse) only goes linearly with velocity.
Since no process can be 100% efficient, you're dealing with quadratically more and more waste heat from the engines. And space is an excellent insulator.
Well, if you create a nice way to make and store anti-matter, you would basically solve rocketry on any place that humans could ever care about going to.
There are people researching that too, but not as many as fusion. And well, determining the capacity of things that don't exist is the first step on the work of making them exist.
For all Musk's faults, he recognizes such timelines are untenable, and pushes people to do the 'impossible'.