Hacker News Reader

Comment by _fizz_buzz_ | Hacker News Reader

A counter point by Sabine Hossenfelder: https://www.nytimes.com/2019/01/23/opinion/particle-physics-...

I post this every time someone links to Hossenfelder, but please take her opinions with a grain of salt.

She seems more interested in attacking the "physics establishment" than in giving people a picture of the reality of our situation. Take a look at her articles on LIGO [1] for instance where she makes conspiracy theory level claims saying that the group is lying about their data and that "big physics" is in on it.

My opinion happens to be similar to hers on building the next big collider, but I'm not a high energy physics expert. Given how dishonest she was on the subjects I am familiar with, there's no way that I trust her to give an accurate portrayal of the ones I'm not.

[1] https://backreaction.blogspot.com/2019/09/whats-up-with-ligo...

I like her videos because she's refreshingly honest about what is known versus not known. Most (tv)physicists are not.

For what it's worth, I usually find Sabine Hossenfelder to be insufferably critical and self-promoting, but I found that LIGO article fairly disturbing. As someone who has worked with complex analyses of large datasets, I know how easy it can be to fool yourself.

Certainly, and the LIGO is one big complicated dataset.

The problem is that Sabine Hossenfelder doesn't just make arguments about the data, but makes the researchers out to be villains. Not only that, but all of the problems she brings up have good community accepted resolutions that she selectively ignores.

I think that once we had multi-messenger confirmation of merger events, any doubt about gravitational wave detection vanished outside of some fringe groups. All of that happened and played out long before Hossenfelder's article, but she doesn't really give a fair representation of it.

Not to defend her, but to defend listening to her...

Often the people in a position to give the most important feedback are _not_ in a position to give feedback gently and tactfully.

So if you want to really know the world, you need to have a check in your mind that lets you take tactless feedback.

One example is when you hurt someone... a person you hurt is almost never going to be in a position to gently, calmly say what happened in a fair, generous way. That kind of feedback almost always comes with barbs, or inflations that accompany a person trying to talk to someone who hurt them.

If you ignore tactless feedback, you will mostly never be able to understand the people you hurt.

Another category like this is feedback about norms. People who are well versed in norms and good at operating within them generally won’t be the people who can see outside of those same norms. The people who can see alternatives to norms, who can give feedback on where they work and don’t work, often will not be able to give that feedback in a way that conforms to other norms.

So if you ignore tactless feedback, you will mostly never be able to understand broken norms.

If you can separate the content of the message from how it’s delivered you can dramatically increase your access to understanding.

In a scientific sense, NOT finding something should be equally valid as finding something. Just do proper work and write it up and we have increased our understanding of the world. If there was nothing, not even the Higgs, so what? There is nothing to defend or to attack. Where is this discussion coming from?

The actual timeline for this is that when the first critiques of LIGO's analysis came out, there was a good reply from LIGO and a productive scientific discussion, at the end of which Sabine indicated she was satisfied. That was months before. That was good.

Then she came out with the linked post, which pretends LIGO had no reply to the substantive issues even though she already had blogged positively about their reply earlier, and disingenuously throws in a few complete non-issues just to make a gish-gallop. That's not science popularization, that's spreading FUD for the sake of self-promotion. Any field can be "taken down" with enough dishonesty.

In the late 19th century people that went into science where actively discourage to study physics since many believed there was nothing left to find out. Without quantum mechanics we would not have the transistor, there would be no computers. The immense economic value delivered by fundamental research makes a 10 billion usd investment look like a microscopic peanut crumb compared to the potential value it can give. For comparison the combined world defense budget in 2018 was around 1.8 trillion usd.

Difference is, the discovery of quantum mechanics did not require billion-dollar investment in equipment and manpower (not until work began on nuclear weapon, anyway). Besides, current research in high-energy physics, while being expensive, only brings scientific value and not much at all in the way of practical application (except, perhaps, as a side-effect of building more and more sophisticated equipment - but that would be like saying that wars are good for medicine, because obviously direct investment in medicine would be more efficient than investment in war).

How can you be sure that the new science won't have any practical applications?

We haven’t seen any evidence of that in the many years that have passed. We have good science, the Higgs and all, but, frankly, there’s little hope that the energies comparable to those that existed at the beginning of the Universe would be of practical consequence to us (other than the destruction of the Earth).

If they do it’s nothing we’ll ever be able to harness

Why should society at this day and time then plow millions of man hours and raw materials into proving it literally?

Perhaps when nanotechnology or AI is able to be put to the task, those peoples can then build a very specific machine to prove very specific questions

But this is throwing spaghetti at the wall for what to us will be a moment of excitement then having no clue how to make use of it and a big mess to clean up

>> How can you be sure that the new science won't have any practical applications?

Nobody is sure about anything. Nobody denies that high-energy physics might one day bring great things. The problem is cost. Other areas of science are much more likely to bring great things with far less money. If tens of billions are to be spent on a bigger microscope, let it be one that will give us better solar panels, or a direct treatment for viruses. Polishing the details of the standard model can wait.

This is just burden-shifting, and if it is the best argument that can be made, there would be essentially no case for a bigger accelerator at this point.

Things are not quite that bad, but the case is weaker now than it has ever been before.

I think its too dismissive to claim it will have 0 practical applications. Often this type of fundamental research take decades to result in something useful in an unexpected way. We can only go where the roads lead us, who are we to dismiss the roads not taken.

You're dismissing other roads not taken.

What is more likely to revolutionise physics - spending $20bn on an incremental increase in collider energies, on spending $20bn funding a new generation of PhDs and postdocs exploring quantum gravity fundamentals?

At this point practical HEP shows every signs of being a boondoggle. There are plenty of hard theoretical questions that haven't been answered, but with a few exceptions they're outside of the mainstream. Research into them has been actively discouraged, except at a few locations.

Physics doesn't need more hardware, it needs more ideas - more intellectual diversity, and more creativity.

I'm not sure just increasing funding to thep would generate new ideas, why would they not just funnel it into the same "safe" programmes as they have done the last 25 years? There are some small "fringe science" programmes here and there, but we probably shouldn't pour all money into those either..

I was under the impression that physicists had too many ideas and too few chances to test them in experiments.

The method itself, smashing ever more energetic particles into each other, makes it highly unlikely that any discovery will be applicable to human sized energy regimens.

Current high energy physics research brought us lots of big data advances, electronics advances, sensor advances, as well as the web itself. You dismiss too quickly the advances made when attacking large, hard problems.

Have you considered some of the large, hard problems whose solutions also have immediate benefits?

Yep - but that is not true for most high energy physics. The incredibly short lifetimes and low coupling of most new particles make them mostly unusable for anything we can forsee being useful.

Electrons were useful, because they have long lifetimes, are easy to make, and interact well. Higgs bosons are unlikely to be useful as technology ever since they are so ephermeral. And they're certainly not giving immediate benefits.

Why kid ourselves that “we are doing physics,” then? Wouldn’t direct investment in these areas more efficient?

>Wouldn’t direct investment in these areas more efficient?

Often not, because you don't know what areas to choose. Having a big, hard goal to solve usually requires building pieces to solve the problem, and these pieces end up being useful in themselves. Plus the big goal is often achieved.

Mankind could have done a lot of the stuff discovered along the way to big goals, but didn't, until the big goals were attempted.

Plus, it's often easier to get funding as a nation-state for big goals. Little side projects are often not individually worth chasing. The internet is a good example - it could have been done by industry before the govt decided to chase it as a piece of a big goal.

I agree and the enormous amount of knowledge build up is lost overtime when not actively used. Not investing in the next project is like no more space travel after we went to the moon. We should always push these projects further for the sake progress. Especially when you see it in the light of the defense budget.

Can you please elaborate on QM being necessary for the transistor? Was it necessary for its invention in Bell Labs?

It isn't that we wouldn't have transistors but that we wouldn't understand them. We wouldn't have the good transistors necessary for modern technology.

There is a similar debate around Newtonian physics. How far might the industrial revolution have gone without a perfect understanding of how things move? You don't need newton to build complex machines. Look at any early windmill or waterwheel. You need newton to understand machines well enough to build and run them efficiently. You don't always need theory before practical, but theory always makes practical better.

You can build a windmill without Newton, but he makes it much easier!

Newtonian physics lets you work out loads with free body diagrams, derive Euler beam theory, and that's how they built the Eiffel tower.

I found a Physics Stackexchange thread addressing this.[1] The science of quantum mechanics was not necessary for its invention. I’m still uncertain of its direct application other than “at <32nm QM effects are dominant”.

[1]https://physics.stackexchange.com/questions/112615/why-is-it...

Hum... That thread goes a long way about diodes, but I couldn't find any such claim about transistors.

Yeah, somebody would probably invent them at some point. Somebody would probably even invent FETs, even if it took many decades more. MOSFETs are a completely different matter, I don't think anybody would invent them without understanding how they work.

What everybody seems to forget is modern chemistry, material engineering and medicine. The world would be a completely different place without quantum mechanics. Electronics is just a tiny part of the change.

MOSFETs are simpler than bipolar transistors. Putting it very crudely, MOSFETs are based on the same space charge principles as vacuum valves, except they happen to use doped silicon instead of free electrons.

Bipolar transistors are more complex, but they still work by controlling the bulk geometry of moving charge.

None of this requires QM, except in the very simple sense that you need to know what a bandgap is and how doping changes it. Beyond that, the "mechanical" details of the charge dynamics are just calculus - even more crudely, very very small plumbing.

QM effects are only relevant when the simple space charge models start to break down at very small geometries. Your "mechanical" space charge model becomes soft, noisy, and more complicated. Essentially the plumbing develops waves and ripples that depend on the geometry, and the edges of the pools and pipes can start to leak. And that's where QM finally starts to make a real difference.

So it really is where the particle phenomenon breaks down, thanks. Do you know if there are any kind of harmonic oscillator models used in this domain? I’m curious about a frequency-dependent characterization of the “energy” transport.

But if you can figure out diodes, you can figure out PN junctions. Well, a bipolar transistor is just two PN junctions. So the same claim should at least somewhat apply to transistors.

No ,you are wrong. First of all we would still have vacuum tubes. Secondly, FETs (their physics is simple, and does not even need that much of quantum mechanics) were invented much earlier than BJTs (the quantum ones) and very possibly we would have been no more than 10-20 years behind from where we are now.

Was QM necessary for invention of bipolar transistors?

It will be impossible to get an extra 10 billion dollars invested into science without a project like this. In other words probably 10% or less of that money would end up going into other fields of research. Considering all the other things governments spend money on, that is a good enough reason to do it in my book.

It’s interesting to see capitalism’s influence here. Like in many other areas of the economy, capitalism requires a concentration of control in science too, as a prerequisite for allocation of capital. Trapping us in a local minimum of narrow hierarchies.

What do you mean with "the other things governments spend money on"? That's infrastructure, education, welfare. The 10 billion is not free money but taken from taxes, so please make a good case for spending it! Surely any intent of not being honest about the cost/benefit will backfire and discredit future scientific projects.

Military spending (eg USA, Russia, China), politically motivated public works (eg Japan), propping up failing industries for political reasons (eg car industry in Australia or USA), not to mention all the money that gets wasted on consultants, starting projects and not finishing them, poorly thought out stimulus packages...

There is a huge double standard where some things undergo a stringent cost/benefit analysis and other things are not scrutinised at all. Also the upside or benefit of scientific investment can be astronomically and unpredictably high. How can you compare building more public housing to inventing the semiconductor industry?

> I still believe that slamming particles into one another is the most promising route to understanding what matter is made of and how it holds together. But $10 billion is a hefty price tag. And I’m not sure it’s worth it.

When is it worth it? They tested many theories, and many of them were proven wrong, and some were proven right. What price can you attach to having this knowledge? Is there a maximum price tag for potentially knowing more about the universe?

The question is, what kind of other interesting knowledge might we acquire with ten billion?

> Is there a maximum price tag for potentially knowing more about the universe?

Yes. On the extreme end, there’s a finite amount of useful work that can be done before the heat death of the universe. An energy budget the size of the combined output of all power stations on the planet is much smaller than that, but also too much. At a more realistic scale, human knowledge is only valuable in so far as it improves the human condition in some way, now or in the future. Other research areas could be more important to progress right now and we only have so many scientists available to do the work. Allocating those scientists to research areas is an incredibly hard task, but it’s a necessary one.

Even if we accept that pure physics knowledge is a worthy goal in itself, diving straight into another megaproject isn’t necessarily the best way forward. Maybe it’s more valuable right now to find uses for the great theoretical leaps we’ve already made, for instance. Or maybe we should be focussing on training theorists so that we can have a better idea of what the next piece of giant scientific equipment needs to be. Or maybe we need to raise the global literacy rate so that we have more scientists available.

> Other research areas could be more important to progress right now and we only have so many scientists available to do the work. Allocating those scientists to research areas is an incredibly hard task, but it’s a necessary one.

That's not really how it works. People work on things they're interested in, not what they get assigned to.

>People work on things they're interested in

That's not how reality works out to be.

That’s... a bit to simplistic of an understanding.

As a researcher, there are many problems I could find interesting and get passionate about working on, either with the right colleagues, funding, world implications, etc...

The scientist oblivious to the world and only interested in their little niche is not the majority. Give a problem funding at the early stage of careers (grad school, new tenure grants) and it gets worked on and some people get super passionate.

If you're talking about allocating existing scientists, it's a bit late for funding something at the point where someone is deciding what to do in grad school (or even undergrad, in the European system).

kd5bjo seemed to be talking about allocating people to entirely different fields which they likely have no existing experience in.

We’re in a capitalist system, so I had assumed that allocation would happen via the free job market: people with money write job descriptions for scientists to work on particular things, and the scientists decide which of those offers to pursue. Central control of this sort of thing hasn’t historically worked out well (though I may be biased on this point from growing up in the US).

As for switching fields, one of the key parts of scientific training is how to learn things, so scientists have a better chance of successfully changing felds than many other people. It also isn’t necessarily as big a leap as physics to biochem, like you suggested in another comment. Moving from theoretical particle physics to tokamak design seems like it should be doable, for instance.

People are "interested" to work on things partially due to the fact there is money available to work on those things.

> That's not really how it works. People work on things they're interested in, not what they get assigned to.

You have no idea how management and funding of large-scale research projects works.

Indeed, government-funded national labs are highly bureaucratic institutions.

At the level of picking a field, people absolutely work on what they are interested in (and have expertise in). You're not going to get particle physicists working on organic chemistry just because you've "allocated" them there (and they wouldn't be trained in it anyway).

> You're not going to get particle physicists working on organic chemistry just because you've "allocated" them there

If I consider what kinds of crappy jobs some excellent physicists who could not find a position in research had to take, I am not so sure.

The truth in my opinion thus rather is: Because, as you wrote, "they wouldn't be trained in it anyway", they won't be allocated for "organic chemistry" in your example.

On the other hand, if they were, I really believe that they would prefer this job over the one that they had to take.

There's not really a million [1] lawyers because all million have a passion for corporate law, its also because it pays well

[1] https://www.americanbar.org/content/dam/aba/administrative/m...

> human knowledge is only valuable in so far as it improves the human condition in some way, now or in the future.

We have no idea how it will improve the human condition though. You could have said this to many scientists throughout history. Look at what the (much cheaper) fooling around at Bell Labs got us.

You’re absolutely right, and those projects were absolutely worth the cost to society in retrospect. It’s why I’m generally in favor of exploratory research.

I don’t have much of an opinion one way or the other about building another particle accelerator: the costs are significant, but manageable, and I haven’t followed the science enough to get a good idea of the likely benefit.

The question posed, however, was whether the cost could ever be too high. The only answer to that question is “yes,” regardless of the topic.

Something that really sunk in for me while arguing with people on HN is that the entire Nuclear power industry is ~60 years old and provides around ~10% of humanity's electricity [0]. Even after the effective moratorium from 1985.

It seems very unlikely that we've really run up against the limits of what physics can do for us. That is an enormous change and it isn't even a lifetime ago. One more development like that would trivially justify investment values in the trillions of dollars range. Let alone all the other stuff that is likely to happen along the way.

Theoretical limits are all very well, but I strongly suspect the rational amount to spend on it is "everything we can". If I really wanted to start a fight I'll argue that diverting all welfare spending to scientific research would probably result in a greater net good over two generations.

[0] https://en.wikipedia.org/wiki/File:Nuclear_power_history.svg.

You are talking about engineering there, not science.

Science is using duct tape to explore the properties of graphene. Engineering is building a plant that can produce 500km2 of graphene a day.

> If I really wanted to start a fight I'll argue that diverting all welfare spending to scientific research would probably result in a greater net good over two generations.

Why welfare spending? Why not corporate subsidies? Or just forcefully seize the entire assets of some number of non-scientists, chosen at random? Cutting out welfare as a whole seems to be the single spending cut most likely to kill the largest number of people.

> Cutting out welfare as a whole seems to be the single spending cut most likely to kill the largest number of people.

Yes. I think the GP's point was that even on the worst case it would still pay off.

Personally, I doubt it. Increasing social risks would move things the other way around and decrease the speed of science advances, even with high investment.

Yes, this is the grandiose technocratic elite who believe they can save the world. A bit exaggerated, but I find these kinds of comments scary, misguided and naive, as well as very revealing of the underlying ideologies and oversimplified & mechanistic views of society.

> I strongly suspect the rational amount to spend on it is "everything we can".

By saying “everything,” you’re dodging the question of what sacrifices we should make to speed up basic science research. How do you draw the line between what we can and can’t afford?

Should we require everyone to pursue a Master’s degree? A PhD? Do we shut down university humanities departments? Do we halt all production that doesn’t directly support the basic sciences? Do we reduce food production to only that which is needed to support the scientific staff?

Some or all of these suggestions are probably beyond your idea of what’s acceptable. If so, you agree that there is some price that is too great to pay. The interesting question isn’t whether there’s a limit, it’s what it is.

>Or maybe we need to raise the global literacy rate so that we have more scientists available.

Or maybe we start paying scientists living wages so anyone who can wipe their ass doesn't end up as a quant. Speaking as a quant with a physics phd working with another 400 phds in an hft firm.

The financial incentives are seriously so misaligned. Likewise as a physics undergrad who considered going into academia and is now doing tech.

I don’t have a day go by without thinking about doing something worthwhile with my (particle) physics phd. But that ridiculous combination of publish or perish, lousy pay and teaching not to mention the competition to stay in the field. Nah, I’ll choose living instead. But oh how I dream of a real career in physics with a higher goal than second-order marketing revenue.

Would you sacrifice all humanity for that knowledge? Since I’m guessing your answer is “no”, then there’s obviously a maximum price (even if my bound is ridiculously off)

That is only 1/300 the cost of the covid stimulus. it is only extensive because it has been deemed by the media and pundits as so.

The issue is, quite literally, what is the opportunity cost of this expenditure. If that money could help create a pandemic mitigation framework that would enable us to collectively respond faster and better to outbreaks, which is better to fund? If the money can help us build faster/more computers that enable us to better model many processes: weather, disease, materials, etc. is that a better expenditure?

I don't see anyone arguing that science is not a good investment.

I do see people asking the salient question of, is this the right science to fund right now, given all of societies needs.

The answer to that question should be obvious, we have other priorities at the moment.

If you can stand his idiosyncrasies, Lubos Motls offers many direct counter points to Sabine:

https://motls.blogspot.com/2019/03/some-reasons-why-west-won...

https://motls.blogspot.com/2018/11/new-veins-of-science-cant...

Reply on news.ycombinator.com