Hacker News Reader

Comment by limaoscarjuliet | Hacker News Reader

According to Sabine Hossenfelder, published numbers are frequently mis-stated on purpose. When you look at frequently reported Q Plasma (or plasma efficiency) we are not that far from it being > 1. However, we should look at Q Total (total efficiency), which is still way below < 1, even in the best plans.

https://www.youtube.com/watch?v=LJ4W1g-6JiY

For years, I was hoping fusion is close. After watching Sabine's video, I'm not so optimistic anymore.

View on HN

She makes several significant mistakes in that video. One example is the energy used to heat the plasma is now heat in the plasma identical to the heat generated by fusion. This means you can recover a percentage of that with a steam turbine.

Second a great deal of ITER’s energy usage is as a science experiment not a fusion reactor. Most of their monitoring equipment for example is irrelevant to an operating power plant. Thus Qplasma is giving relevant information where Qtotal is largely meaningless at this stage.

But her point stands even if there were mistakes: the press and many scientists, even if unwillingly, have failed to communicate the real state of fusion as an energy source. And I can believe that maybe some representatives did not fully understand the difference between Qtotal and Qplasma and the amount of time a reaction can be sustained (all three things that they would understand if they were explained clearly).

Some context please. Where did they make this failure in the hour long press release?

The parent comment was talking about a video outside of the context of this entry in HN. My reply was in the context of parent comment referring to that video, the video in question provides excerpts of example of that failure of communication.

Sabine is overly critical IMO. The distinction she is making is known to anyone who has spent a little bit of time thinking about fusion (hopefully that includes the grant writers) and projects like ITER are explicitly aiming for Q=10, not just "breakeven"

Which Q? Qtotal or Qplasma?

Qplasma. Supposedly, with Qplasma > 20 Qtotal should also be > 1 with current magnet tech.

ITER is a bit too optimistic IMHO. For example I've always found the "UNLIMITED ENERGY" on their website a bit funny https://www.iter.org/

That's kind of her thing. She is usually right, but also overly pessimistic in a way opposite from the popular press over-optimism.

I get frustrated because it paints science in an undeserved negative light. It is at least truthful, in a way that most anti science writing is not. Mostly I find it unhelpful in that it points out problems without either explaining why they were reasonable or giving a real alternative.

Popular 'i f#@cking love science's type science deserves to be painted in a negative light.

IFLS pushers are motivated by money, clicks, and clout and to a large extent misinform the public.

Agreed. But there is blowback on the actual scientists doing the real work.

I suppose you could say that they also get benefits from appealing to the IFLS crowd, so live and die by the same sword. I believe IFLS does more harm than good, but it's hard to be sure.

Indeed. Her arguments made in the aforementioned video are slanted. There are many lies by omission that paint an inaccurate worldview for laypeople.

It's incredible that the term "Lawson criterion" wasn't mentioned once.

Sabine is not as "right" as many of her pop-sci fans seem to want. Most of her opinions about the direction of theoretical physics are not really falsifiable or "right."

That's correct. That's sort of the point. Fundamental research is in kind of a slump right now, and it's hard to judge where (or whether) it should continue.

I don't think much of her "just do something different, don't ask me what" approach.

Nothing concrete to add, just the anecdote that the plasma physics class I took in grad school hands down had some of the sketchiest looking physicist math I think I have ever seen. Felt like a SWAT team from the mathematics department might burst through the door at any time.

Like an ECON 101 course?

I felt like the Calculus SWAT Team was going to burst through the doors in that class, ha.

This was no regular spherical cow, I tell you it was assumed to be time-independent while at the same time having an oscillatory frequency.

Just imagining the mathematics SWAT team coming through the door and swatting the marker out of one of my physics profs' hands when they went from ydy/dx=x -> ydy=xdx as if they were simply re-arranging a fraction, made me chuckle.

Why would any experimenters optimize for Q total if Q plasma is a prerequisite to get anywhere?

The problem is press/marketing. When you read about this stuff 1.0 is claimed to be break-even, the threshold where fusion start to become practical. So in one piece of writing they will talk about this important number and (deliberately) conflate it with fusion being practical. Yes, getting Q plasma above 1 is necessary, but it's about an order of magnitude too low in reality. Sure everyone in the field know that. Her criticism is thats not how its presented to the outside world.

The concern is that experimenters are wasting time on experiments with promising looking Q plasma but with orders of magnitude smaller Q total (eg pulsed laser systems)

It’s important not to forget the big picture. Otherwise you end up optimizing one piece of the system, and causing another piece of the system to work less well in a way the degrees overall system performance.

Well, because it makes no sense to try to increase Q plasma; to what end? As pointed by others, you can increase Q plasma at the expense of Q total, thus precisely optimizing for the wrong objective (fusion can be produced and studied without looking at Q plasma or Q total; although reporting Q plasma along with the time that the reaction was sustained is helpful). I give you an analogy: let's make the most power-efficient floating point unit; but let's focus on Q_float, instead of Q_total; at the end we would end with a simple very wide adder; if we want to multiply two numbers, the system will convert them to their logarithms, add them, and then use exponentiation. Yes, the floating point unit consumed very little power Q_float was great. Well, I guess you get the point.

Reply on news.ycombinator.com