Hacker News Reader

It's premature to say what's practical with fusion and what isn't. It's entirely possible that we find a way to build an efficient hydrogen fusion reactor. In which case it might not make sense to travel so far away from our planet just to collect energy from photons when that costs more than just doing the fusion ourselves. Remember it's actually more expensive to travel towards the sun than away from it (because you have to shed velocity relative to the starting velocity on Earth.) It's too arrogant to think we know the answer to that now.

> It's premature to say what's practical with fusion and what isn't. It's entirely possible that we find a way to build an efficient hydrogen fusion reactor.

Would it not surprise you that fusion in general has been studied extensively by physicists for nearly a century? It takes very high temperatures, pressures and quantum tunneling to get two protons to fuse.

Here's some reading to get you started:

https://courses.lumenlearning.com/physics/chapter/32-5-fusio...

Or read about how the Large Hadron Collider works.

It doesn't surprise me. Your arrogance does though, I didn't expect you to double down on an unreasonable position. How can we possibly know what we don't know yet. Not about how fusion works in theory, but about the limits of the economics of fusion reactors at any point in the future.

What I'm saying is it's too soon to say what's possible with fusion and what isn't. We don't even have a break even reactor yet. We can even say for certain that we ever will.

> It doesn't surprise me. Your arrogance does though, I didn't expect you to double down on an unreasonable position.

I'd give your criticism some consideration if I thought you had even the slightest knowledge of what we're talking about.

Let's talk about the core of the Sun, where p-p fusion takes place. Remember, I said at the top: no new physics.

The core of the Sun is hot, like really, really hot. How hot? 15 million Kelvin. So no solid material ... made of atoms ... can be anywhere close by because it would vaporize. You can talk about active cooling or whatever, and that's not going to help. But let's assume we have some magical material that can deal with that temperature, and keep going.

We're going to skip the pressure that really hot plasma is under, and assume we have terrifically awesome magnetic confinement that can handle that. We don't actually, but let's move on.

The other problem with 15 million K is that all that energy wants to radiate away. Yes, you want to absorb some of that radiation (hard gamma, by the way, very nasty) as part of the energy generation process, but if you let the plasma cool off, the reaction rate drops. "Cool" in this context is relative, at 7 million K, the reaction rate is very close to zero.

The last problem we will talk about today is volume. The above conditions get us a fusion power density of... 276 watts / cubic meter. To have a fusion plant that generates 1GW of power, we'd need to contain a cube 153 meters on a side of 15 million Kelvin plasma.

Oh, wait, that assumes the magnetic containment and everything else requires zero power to run. So we'll have to go bigger, much bigger.

Are you starting to understand the scale of the problem... trying to design a proton - proton fusion reactor?

I am not the arrogant one, I'm just listening to the fusion scientists who have been studying how all this works.

> We don't even have a break even reactor yet. We can even say for certain that we ever will.

And here you are expressing doubt that we'll achieve even the easiest kind of fusion reactor. We have achieved fusion, just to be clear. We can reach the temperatures and pressures to see some reactions.

Assuming civilization does not collapse (climate change, war, etc) we will be able to engineer a fusion reactor that is some kind of practical.

But it won't be proton-proton.

Edit: Actually, the material issues mentioned above probably won't be an issue, as long as the confinement is good. You still have the issue of keeping the plasma hot enough for long enough.

But the scaling issue doesn't go away unless you go with even higher temperatures and pressures.

> I'd give your criticism some consideration if I thought you had even the slightest knowledge of what we're talking about.

Doesn't sound like an arrogant person to me...

FYI I understand how magnetic confinement fusion works at a high-level, as a non-scientist.

There's a lot of assumptions you're making here. First that we need hydrogen-hydrogen fusion. Many current reactor designs use deuterium and tritium (nasty stuff tritium.) It doesn't have to be plain hydrogen, it just has to be elements common enough that we won't run out (or that we can breed using neutrons - that gamma radiation you talk about - from the fusion reactor.) Some designs are more exotic with fuel using isotopes of helium, lithium, boron. Second many current reactor designs use a plasma temperature well above 15M K. I've heard of 100M K and that's not an upper limit.

But you're also making assumptions about the volume of a reactor - we don't even know if magnetic confinement fusion will end up being the best design, but if we assume it does - then stronger magnetic fields really decrease the volume required. In fact development of stronger magnets based on high-temperature superconductors may be the biggest breakthrough in fusion research in the last decade - see https://news.mit.edu/2021/MIT-CFS-major-advance-toward-fusio...

> And here you are expressing doubt that we'll achieve even the easiest kind of fusion reactor.

We don't have a break-even reactor yet by the way - we have achieved fusion for short periods of time and by putting in far more energy than could be produced. That milestone is probably still at least a decade away.

My point again is that we don't know enough to be able to say what's possible and what isn't. History is littered with people arrogant enough to say what we can never achieve and humans keep surprising. Sometimes even things that seem impossible - like detecting planets around other stars when the physics of optics is very clear that an optical telescope could never directly image a planet in another solar system. That was wrong on two fronts, one it doesn't account for gravitational lensing, and two it didn't account that we might cheat and detect planets indirectly. And so the arrogant people were wrong again. Don't make that mistake.

> and if you want to collect _all_ the energy from a star,

For what? What are you doing with it? Not to forget the fact that all energy used eventually turns into heat. All of it (with the exception of electrochemical transformations). You're bringing all of that into Earth you're cooking everybody alive.

How are you bringing that energy from the sphere to (wherever place)?

Putting solar panels on Earth itself works much better. Heck, make a Dyson sphere on Earth (facing outside), that makes more sense.

It is typically assumed that by the time you are making a Dyson swarm, that most life exists digitally. And spread out through the solar system.

So in the best case scenario, you are powering computronium, upon which trillions of sophonts live out their digital lives. Hopefully doing some science, or at least playing some interesting games. A vast distributed network.

In the worst case scenario, Blockchain and spam emails.

Either way, people will put that energy to use.

With a Dyson swarm, energy is received on the inside, towards the star, and heat is radiated outward. Lower power systems can run off the heat radiation of the inner swarm similarly. It is all eventually released as very low level heat, but diffused across the entire outer surface of the swarm. No hotspots... because someone would come along and use it.

The Earth would also likely be deconstructed by this time, so don't worry about the biosphere, there won't be any.

> For what? What are you doing with it?

One example: powering lasers for light sail ships to colonize other star systems.

You are not pushing a light sail ship to another star system, sorry.

Light sails are "theoretically" possible but you don't need that much energy to push something that's feasible to push.

Or better, there's a better source of light radiation: the sun itself. No need for a laser, just use a focusing mirror and you can have 100x more pressure than just from the sun itself.

Light sails are practical, and have already been tested in orbit.

https://www.planetary.org/articles/lightsail-2-successful-fl...

Light from the Sun isn't coherent, and you can't focus it with mirrors very far away. Lasers have many orders of magnitude less divergence when talking about distances measured in light-years.

> Light from the Sun isn't coherent, and you can't focus it with mirrors very far away. Lasers have many orders of magnitude less divergence when talking about distances measured in light-years.

To expand on your point, those characteristics become even more important when using the light received to decelerate at your destination, though it is worth noting that there are alternative means of deceleration that don't have the same requirements (eg. electric and magnetic sails).

Why not? Light sails can reach 10% the speed of light, and sending out thousands of colony ships in multiple directions seems like a feasible way to colonize a galaxy relatively quickly.

As the other poster mentioned, harnessing the sun's energy and turning it into a better form suited for light sails is the way to go. Sailing on the solar wind within a star system is a different tech with a different purpose.

> Heck, make a Dyson sphere on Earth (facing outside), that makes more sense.

I’m missing something here, because even if we decided to build a Dyson swarm of course we are going to start with putting solar on and around Earth because this is where all the stuff we care about is.

That’s just the first step in a journey of a miles miles. (TIL where that quote comes from: https://en.wikipedia.org/wiki/A_journey_of_a_thousand_miles_...).

This seems so obviously it can’t be what you’re taking about though.

> We have two basic ways of generating a lot of power, fission and fusion

I think there are more. Like radiation from evaporating black holes.

> I think there are more. Like radiation from evaporating black holes.

Evaporating black holes?

Yes, if you are willing to wait a very, very long time, and have collectors over a large area.

However, if you want to get stuff done before your protons evaporate, you could instead throw matter into a black hole, thus increasing its mass. The radiation given off by the accretion disc is substantial, and can be collected... over a large area.

IIRC, the mass-to-energy efficiency is something like 40%, which is really good.

They are a little difficult to set up however...

I mean, we're talking about timescales and construction projects like Dyson spheres. It's entirely feasible that people with such technology could create microscopic black holes that evaporate quickly for these purposes. Hawking radiation at this scale is substantial.

Ah, yes, I do recall reading about that somewhere.

That is an interesting prospect, with a very high mass-to-energy efficiency. Of course it requires careful management of inflowing matter. And you can shut if down if needed by stopping the inflow, assuming you have enough stored energy to recreate the micro black hole sometime later.

You'll probably enjoy this video:

https://youtu.be/EzZGPCyrpSU

Good channel too.

A Dyson sphere is technically possible without any new physics with a few decades of work. By contrast, while we have the physics to analyze how black holes much smaller than a proton would behave, we have no plausible way to go about making one.

> A Dyson sphere is technically possible without any new physics with a few decades of work

Not a chance. I disagree 100%. The materials science and orbital mechanics for coordination alone are way beyond our capabilities.

Dyson sphere as conceived by Dyson himself is a swarm of orbiting objects, not a literal sphere around a star (that's Star Trek interpretation). We could start sending out solar panels to orbit the Sun today, and start adding space stations later this/next decade.

It doesn't count as a Dyson swarm until it captures a large percentage of the sun's output. As I said, that science and the computational power needed to coordinate those kinds of orbital mechanics are beyond us.

You'll see.

It doesn't all have to be at a single distance from the Sun. Some objects can be as far as Earth is and other objects can be at Mars level, at Merkur level or wherever else. There's plenty of space for this to be too hard.

That doesn't matter if we're talking about scenarios where a significant fraction of the sun's surface area is covered. That requires enough mass for each fragment to make such a loose "shell" that they would influence each other gravitationally and perturb each other's orbits.

Not to mention that that much surface area would catch a lot of solar wind, causing all sorts of rotational moments that will need constant correction. The control theory needed for this sort of thing is complicated and N body problems are undecidable in general.

It would catch approximately all the solar wind. "All sorts of rotational moments that will need constant correction" means that, if your orbiters are thin enough, they don't need reaction wheels or thrusters to change course to avoid an imminent collision — just solar-sail louvers. An occasional collision is unimportant to the viability of the system, like dust settling on solar panels, or dermis cells dying in a follicle to grow hair.

N-body problem instances are never undecidable in the automata-theory sense; you can answer any question about the future trajectory from given initial conditions straightforwardly, if slowly, with interval arithmetic. The question of whether initial conditions exist that evolve to a given result is undecidable, but as with undecidability results in general, we do not need to decide it in order to build a working system, only to prove the optimality of a given program.

You seem to be making up reasons to justify a preselected conclusion without bothering to think about whether they are true or relevant or not, and I wish you would not do that.

Not significantly. Do the math: a thickness of 100 microns with existing silicon solar panels, or perhaps as little as 0.1 microns, out of, say, 1 AU (1.5e11 m). We're talking about filling 2e-15 of a 1-AU sphere with solid orbiters, totaling roughly the mass of Earth's moon, if we're using current silicon solar cells, and potentially a few orders of magnitude less than that.

You could use a smaller or larger radius, resulting in higher or lower irradiance and higher or lower orbiter density, but we have a lot of experience with different kinds of solar panels at 1 AU, so we know it will work.

No, there's no problem with orbital dynamics, control, computation, communication, or any of that nonsense. 50 years ago those were potentially challenging problems, but not now. Today the only remaining obstacles for a Dyson sphere are manufacturing and politics.

It's just a bunch of automated drones flying in formation. Trying to calculate their orbits ahead of time in any kind of detail would be a fool's errand, instead we'd just use solar light pressure to steer them as needed (they would be very thin, there's no point to making them thick).

With really thin sails (0.78 g/m^2) they don't even need to orbit, as they can use light pressure to compensate for gravity (what Robert Forward termed a 'statite'). The light they reflect hits other sails on the other side, but this isn't a problem if they are distributed uniformly.

I wonder how practical it is to sustain right-sized artificial black holes. Small ones should irradiate much more.

Existing proposals use a plumb bob mechanism or directed light traversing the ergosphere in analogy with sound amplification via linear or in this case rotational energy transfer rather than relying on the feasibility of a complete sphere to absorb any randomly directed radiation from the interior or accretion disk: https://news.ycombinator.com/item?id=30402679