Hacker News Reader

Comment by native_samples | Hacker News Reader

Do we need more physics research? I guess it depends a lot on how you define physics, but it seems like outside of possibly better silicon nodes, there isn't a whole lot of low hanging fruit in physics at the moment. When physics does get research funding it gets dropped into building giant machines that, at vast expense, have discovered virtually nothing.

It feels like right now most areas of academia are consuming far more resources than their useful output could really justify, which is perhaps why in so many fields it's so heavily dependent on government funding (vs say computer science where academic/corporate lab collaborations are quite common).

View on HN

It is really hard to predict when research will become valuable- the most valuable US Government grant ever just might turn out to be the NSF/DARPA grants for information science management that funded Brin and Page in their initial work on the PageRank algorithm.

But my favorite example of how research flows into a better world is the Sagnac effect. In 1913, French Physicist Georges Sagnac built a circular interferomter and found interference bands. He thought that this disproved Relativity and showed that an aether existed, but it turned out that German Physicist Max von Laue had predicted the existence of those interference bands under relativity two years earlier, so 10 points for Einstein.

For the next 50 years Sagnac interferometery was a dead end, a minor curiosity in the history of physics. Then in 1963, Macek and Davis at the Sperry Gyroscope Co. figured out how to build this in a laboratory environment with the recently invented lasers. The coherent beam of a laser unlocked the usefulness of the Sagnac effect. Meaning that just another 30-odd years of work by hundreds of people around the world got to a situation where ring-laser and fiber-optic gyros are superior to mechanical gyroscopes and capable of things that mechanical gyros could never do.

So, the Sagnac effect itself was worth nothing, and for a long time afterwards was just something that a few scientists even knew about. But a century later the world depends on it.

"There is nothing new left to be discovered in physics now. All that is left is more and more precise measurement." -- Lord Kelvin, 1900

What is the likelihood of discovering a new phenomena or largely unexplored field of science akin to the discovery of electricity and electrochemistry?

There may be a finite amount of works on physics that we can do, or technological niches that we can exploit thanks to discoveries in physics or any relevant field of knowledge.

The well of discoveries isn't running dry today, but it may in future generations.

I'd say the likelihood seems a lot higher now than it did in Kelvin's time.

> What is the likelihood of discovering a new phenomena or largely unexplored field of science akin to the discovery of electricity and electrochemistry?

We have zero idea how gravity works, how to predict the properties of vast stretches of new materials, if the island of stability is real, what the limits to know propulsion technologies are, what the limits to know fusion technologies are, et cetera. And I’m not even getting to batteries or biology or hosts of related fields.

Everyone here is missing the point. It's not that physics is running out of questions to ask. It's that the questions are getting increasingly more expensive to answer, and the answers are increasingly less compelling. The low-hanging fruit picked long ago, modern physics produces fewer discoveries that change people's lives in the way that radar, lasers, microwaves, and transistors did, which makes science investment less compelling to the public.

There's always a time gap.

Remember that from the first public microwave demonstrations in 01895 by Bose to the first deployments (as radar in the 01940s) took 45 years; the Amana Radarange brought microwaves into people's homes in 01967, another 27 years later; and microwave ovens didn't really go mainstream until about 01990, another 23 years after that, 95 years after Bose's first public demonstrations. The LASER was first built in 01960 (following numerous physics advances starting in 01917) but the first mass-market laser product was the CD player, introduced in 01982 and surpassing vinyl records in sales for the first time in 01988, 28 years after Maiman's first LASER.

If anything, the time gap seems to be shortening, but the place to look for "discoveries that change people's lives" is not in basic physics research of today but basic and applied physics of a few decades ago turning into common practice more recently. And in that case there are a lot of examples, especially if we go further afield from just physics discoveries:

- modern LED lighting comes from the physics discovery of how to make stable blue LEDs (though, for reasons of market failure, most LED lights still last only 3000 hours instead of 30,000);

- modern cellphones, computers, GPS, broadband, and in particular broadband wi-fi come from numerous physics discoveries that have enabled chip feature sizes to continue reducing over the last 20 years;

- optoelectronics advances are largely a question of physics, and without even counting blue LEDs, better optoelectronics have over the last 20 years dramatically improved TV screens, computer monitors, cellphone screens, fiber-optic communication, Blu-Ray data storage, and LIDAR for, e.g., self-driving cars;

- the mRNA vaccine for covid, while a biological discovery rather than a physics discovery, was designed within a few day after the genome was published, and seems to have both higher efficacy and less side effects than previous kinds of vaccines (though unfortunately for political reasons it wasn't rolled out for another 9 months, during which tens of millions of people died);

- lithium-ion batteries have gone mainstream, enabling a transition to electric cars and wireless power tools;

- better power electronics, resulting from solid-state physics research, have made induction stoves widespread;

- modern solar panels cost a tenth of what they did a decade ago, in significant part due to physics discoveries over the past 20 years, now account for the majority of new power generation capacity being built, and will probably dramatically drop the cost of energy by 02030;

- due to chemistry discoveries, Spectra/Dyneema fishing line is cheap, strong enough to make bulletproof vests, in fact as strong as the strongest steel, and floats on water;

- superhydrophobic coatings, a physics discovery, are going mainstream now.

Radar, microwaves, transistors, and nuclear physics (which you strangely forgot to mention, even though it's fundamental to modern oncology, and produces a significant part of the electrical power in many countries) resulted from WWII. We've had an atypically low level of great-power wars over the last 75 years, which has been great, but it wouldn't be surprising to have another great-power war in the next decade. If that happens, maybe the survivors will be reduced to sticks and stones, but if not, you can bet that they will have spent a lot on physics research.

> vaccine ... wasn't rolled out for another 9 months, during which tens of millions of people died

The total death toll has just passed 5 million.

Official one yes. Idk about tens of millions but I’m also thinking 5 million is lower than the true numbers. Though I’m also not sure there is a qualitative difference (as much as it pains me to say).

Those are confirmed, diagnosed deaths. Presumably many people died without ever being diagnosed in places like Andhra Pradesh and Congo.

Yes, but those new discoveries are less fundamental.

One of the reasons we had such an explosion of new theoretical physics in the first half of the 20th century (which gave birth to all the practical discoveries you mentioned) is that, in 1900, there was just so much unexplained weird shit.

Shit like:

1. How do you predict the energy of electrons produced by shining light on a surface? And why does light cause electrons to be produced?

2. Why is the speed of light the same in every direction, to really, really high precision?

3. Why are there spectral bands in the light from stars? And why are they sometimes shifted?

4. Why can current pass only one way through a metal pin poking a semi-insulator?

5. Why does pollen appear to move randomly in the sunlight in still air?

6. Why the fuck is the sky blue? Why does light sometimes act like a way, but sometimes like a particle?

Remember, a lot of this weird shit went unexplained for 25-50-100 years. Then, between 1900 and 1950, smart people came up with theories that explained all the weird shit. So we don't really have so much weird shit anymore because most of it has been explained. Sure, super-conductors are kinda weird, but we also seem to be able to predict a lot of their properties from existing theory.

So, we're really just riding the coat-tails of scientists from 100 years ago. People 100 years from now aren't going to have coat tails to ride.

Less fundamental than transistors? How much less fundamental can you get? There were no transistors floating around in the asteroid belt. Unlike, say, fission reactors, no natural transistor has ever been found.

There's a lot of unexplained weird shit today, too, including the arrow of time, turbulence, dark matter, dark energy, "tuning" of physical constants, solar flares, extreme-energy cosmic rays, ball lightning, the unresolved inconsistency between quantum mechanics and special relativity, structures up to the edge of greatness like the Sloan Wall, the bizarrely low-entropy state the universe began in, and consciousness. Short gamma-ray bursts were finally explained in 02017 thanks to LIGO. (The explanation had been hypothesized previously, but Lucretius correctly hypothesized the explanation for Brownian motion, too, 2080 years ago.) You may be interested in my longer overview in https://news.ycombinator.com/item?id=29144119.

Your list of 6 weird turds unexplained in 01900 is mostly correct, except that #6 was correctly explained by Rayleigh in terms of Maxwellian electrodynamics in 01881, and Fizeau correctly explained the redshift of stellar spectra in terms of the Doppler effect in 01848, though of course not the spectra themselves. We could add, "Why is the Andromeda Nebula's spectrum so smeared out?" (leading to the Shapley-Curtis Great Debate in 01920), "How can the Earth be older than the Sun?", and "What powers Becquerel's uranium rays?"

The quote from Kelvin in this thread turns out to be incorrect (https://en.wikiquote.org/wiki/William_Thomson#Misattributed https://www.bbvaopenmind.com/en/science/physics/lord-kelvin-...) but Albert Michelson did make just such a claim in 01894:

While it is never safe to affirm that the future of Physical Science has no marvels in store even more astonishing than those of the past, it seems probable that most of the grand underlying principles have been firmly established and that further advances are to be sought chiefly in the rigorous application of these principles to all the phenomena which come under our notice. It is here that the science of measurement shows its importance — where quantitative work is more to be desired than qualitative work. An eminent physicist remarked that the future truths of physical science are to be looked for in the sixth place of decimals.

The curious thing is that Michelson in 01894 already knew about at least four of your five weird pieces of shit (the Schottky diode had been discovered but was not yet well known), as well as numerous others. In fact, he was personally responsible for discovering #2 in 01887.

So how could even Michelson make such a grievous error, the same error you're making now?

Well, I guess he underestimated the importance of the pieces of shit that were unexplained.

However, I think that error is less understandable today, though, since among the things we know we know virtually nothing about are the nature of 95% of the mass in the universe and the bulk dynamics of plasma, the state of matter in which we see 99% of the remaining 5%.

There's still the question of whose coattails we'll be riding in 100 years, assuming we somehow manage to survive. The Nobel Prize is imperfect and backward-looking by nature, but it suggests Ghez and Gensel for discovering Sagittarius A*, Penrose for black hole robustness, Mayor and Queloz for discovering exoplanets, Peebles for physical cosmology stuff including dark matter, Mourou and Strickland for petawatt lasers by chirped pulse amplification, Ashkin for optical tweezers, Barish and Thorne and Weiss for LIGO, Thouless and Haldane and Kosterlitz for topological order in matter, Kajita and McDonald for neutrino oscillations, Nakamarua and Amano and Akasaki for blue LEDs, and Higgs and Englert for explaining why things have mass. Those coattails doesn't seem obviously worse than the list from the corresponding years 100 years earlier: Guillaume for invar, Stark for spectral line splitting, Planck for quantum physics, Barkla for X-ray spectroscopy, the Braggs and Laue for X-ray diffraction, and Onnes for liquid helium.

But you don't yet know, for example, the significance of optical tweezers for submicron 3-D fabrication, of dark matter for interstellar propulsion, or of topological phase transitions for sustaining life in the Degenerate Era. (Of course I don't know them either; I'm not a time traveler.) NOVA hasn't even made an episode about topological order yet, so you don't yet understand that today's physics advances are fully as important as those of a century ago.

Consider the innovations evident in The Atlantic in November 01921: https://archive.org/details/walpolebeauty00beck. The Victrola (Edison, 01877), tires with anti-skid treads (Dunlop and Continental, 01904), chemical weapons (book review of "The Next War", p. 10; in theory Playfair's cacodyl cyanide, 01854, but more realistically the Hague Convention in 01899, and then massively in 01914), air war, pewter, eugenics, silk lampshades, the electric chair, aeronautics (not yet understood theoretically, so we have to credit the Wrights, 01903), Haldane's metaphysical speculations about "relativity", automobiles (arguably the Oshkosh steam car in 01878, the Flocken Electrowagen in 01888, or the second Marcus car in 01875), motion pictures (Muybridge, 01878, or Anschütz, 01894, or Le Prince, 01886, or Dickson and Edison, 01891-4), thermostats from the Minneapolis Heat Regulator Co. (Drebbel, 01620), clamp-on electric lamps (based on Edison's 1880s designs), personalized pencils with erasers (Lipman, 01858), Smith-Corona portable typewriters ("seniors' theses MUST be typewritten") (Rose, 01906), prenatal clinics, the welfare state ("if we can tax so heavily for purposes of war without raising a protest...", "The taking over by towns and states...of the responsibility for the care and prevention of tuberculosis...meant..."), kindergarten, workmen's comp ("greatly increased demand for safety appliances"), Prohibition, the traction plough, the Panama Canal, "The terrible world-upheaval through which we have just passed...the great war", the resulting inflation of the pound, passports, the SS Imperator, the Bolshevik revolution, the "roar of the city", buses, the League of Nations, and so on.

The crucial thing about this litany is that not one of this astounding list of innovations owes anything to Guillaume, Stark, Planck, Barkla, the Braggs, Laue, or Onnes. Then, even more than now, applied science rode on the coat-tails of basic science from generations before.

Part of this question is whether there is a limit to human cognition that we can leverage to answer increasingly difficult questions. It isn't whether there are still unanswered questions. So this is truly an issue of diminishing resources and exhausted low hanging fruit. One of those resources is human cognitive ability, relative to difficulty of finding new discoveries.

Just as there are true limits to for example pursuing economic growth from burning finite cheap fossil fuels, "growth" in the forms that pertain to the issue at hand... economic, cheap energy consumption, scientific discoveries, human cognitive ability to solve problems..these can all certainly bump up against at least short term limits, and the issue of cost matters greatly in a given time period because even bringing to bear the actual resource we have to develop new science and technology suffers from crowding out if economic "growth" is constrained because of diminishing cheap energy.

This shouldn't be surprising if one conceives of what consistent and steady compound growth even at small rates results in after a relatively short period of time. Inevitable that limits in many inter related areas will be reached...and either a collapse occurs or a long period of low or negative growth happens with all the sort of conflict that entails (competition within scarcity).

> this is truly an issue of diminishing resources and exhausted low hanging fruit

The fulcrum of disagreement is the power of paradigm shifts. Paradigm shifts change how existing knowledge is interpreted, even by a feeble mind. These are, from what we can tell, randomly distributed. There is no indication that we are running out of them. If anything, their rate of discovery is increasing.

Since paradigm shifts are about prospective, not knowledge, there is no reason to believe they are limited. If anything, our growing knowledge base implies the next shifts will be more powerful than the prior ones.

Well, maybe we won't find anything akin to electricity again; the electroweak force is one of the three fundamental interactions we know affect matter in the universe, and knowledge of electricity as such dates back at least to Plato, 2500 years ago. It was unified with magnetism in 01873, and unified with the weak force in the 01970s. Maybe we'll find a fourth one; maybe we won't. (Gravitational waves weren't observed until 02015, and they solved the mystery of short gamma-ray bursts in 02017.) Here's a list of candidates for similarly fundamental discoveries:

1. Maybe we'll find a fourth fundamental force that we just haven't noticed yet. This sounds stupid until you realize that we hadn't noticed Archaea until 01977 or dark matter (85% of matter in the universe) until 01980 (though the phrase is from the 01930s), and we still know almost nothing about the behavior of dark energy, the existence of which wasn't known until 01992.

2. Maybe we'll find a way to reconcile general relativity with quantum mechanics ("quantum gravity").

3. Maybe quantum computers won't work, demonstrating a flaw in the assumptions of quantum theory in the same way that the Michelson-Morley failure to detect an ether wind demonstrated the flaw in the assumptions of Newtonian physics that Einstein resolved with special relativity. More likely, they will work, and this will change a lot of things; their computational power is still poorly understood. They were originally proposed (by Feynman) as an engine for simulating quantum physics.

4. Maybe dark matter and dark energy don't involve a fourth fundamental force like the strong force, gravity, or the electroweak force, but we still know almost nothing about how they behave. So almost everything about them is unknown. Can we use them for communication, propulsion, computation, energy sources, mass sources, etc.?

5. General relativity hasn't been shown to conserve energy or (equivalently) momentum. Does that mean reactionless drives and perpetual-motion machines are possible, or (more likely) that there's a more subtle symmetry to GR that hasn't yet been discovered?

6. Where does consciousness come from? It's the most perceptually salient phenomenon in the entire universe, but we don't have any convincing account of what it is.

7. We know very little about plasma dynamics. We don't know how to make a usable plasmoid gun, we don't know how ball lightning works (or even if it belongs in this item), and although we know they're a magnetohydrodynamic phenomenon, we don't know how solar prominences are formed, and similarly for solar flares, which accelerate some particles to GeV speeds by means we don't understand at all. We don't know why the solar cycle happens. Coronal waves weren't discovered until 01995. We don't know how to stabilize fusion plasma in a tokamak. We don't know if there are significant magnetohydrodynamic phenomena at scales larger than a star, much less larger than a galaxy. We don't know what heats the corona. This is important because a large majority of the matter in the universe is plasma, and our understanding of it is mostly just empirical, like stamp collecting. We're used to thinking of plasma as an undifferentiated homogeneous continuum like a well-mixed liquid, where nothing interesting happens, quite unlike all our complicated organic molecules (which can't survive in it), but obviously from looking at the sun that isn't the case; we really have no idea about the possible complexities. Is this where the Hercules-Corona Borealis Great Wall comes from? Perhaps more excitingly, if MHD makes stable structures possible in large-scale plasma systems (as it evidently does in the sun), are there analogous phenomena that can occur in a quark-gluon plasma?

8. Forget about magnetohydrodynamics for a moment. We don't even understand regular hydrodynamics. Tao's most famous result (02014) was a finite-time blowup in a version of the Navier-Stokes equation, for which he had to use results from automata theory: https://terrytao.wordpress.com/2014/02/04/finite-time-blowup.... What this means in practice remains unclear (Tao: "In principle, it might even be possible in this case that the speed and the wave number both go to infinity in finite time, a scenario known as finite-time blow-up. Of course, such blow-up does not mean that a physical fluid such as water can exhibit this behaviour, but it does mean that the Navier–Stokes equations cease to be an accurate model for such a fluid in these cases."), but it's clear, if Tao's result can be extended to the real Navier-Stokes equations, that it means we don't have an adequate model for fluids in such cases.

9. Okay, and what's up with the profusion of apparently random physical constants? Could they have been different? Are they different elsewhere in the universe?

10. How did the universe start out with such low entropy? Equivalently (or possibly not, depending on the nature of CP violation), why is the past different from the future?

11. Is spacetime continuous, or is it like a sort of foam? Is the foam size really of the Planck-length scale, or is it much larger, as the holographic principle suggests? What happens when you approach that scale? Fundamental particles like protons are far too big to do experiments like this, but in theory we ought to be able to make black holes that are much, much smaller than protons to do these experiments. (You think NIMBY is bad now...)

12. String theory posits a number of other spatial dimensions. Is spacetime really only four-dimensional?

Quite aside from these fundamental problems, any one of which promises a "largely unexplored field of science akin to the discovery of electricity", there are a huge number of things we can create that don't even require discoveries of fundamental new phenomena like those above. Electrochemistry, to take one of your examples, is extremely underexploited because in most cases we don't know what conditions we have to control in order to make our experiments reproducible, and of course medicine is full of unknowns.

Of course anything could happen in future generations, but there's nothing to suggest that your prediction will happen. Rather the other extreme: the well of discoveries is overflowing today, but may become a geyser in future generations.

Considering the fact that what we don't know is potentially infinite, and what we don't know we don't know is a potentially infinite chunk of that, how could that ever be true? :-)

Well, I guess we could hit a local minima where we need some out-of-this-world resources or research to level up, but let's not be so pessimistic so soon. We've only been doing this modern research thing for a few centuries. You could go back 4-5 generations in an especially long lived family and you'd find someone in the family tree that was still alive before the modern scientific method was created.

In addition to all that's been said about this quote, in my view there can actually be new physics in those precise measurements. For instance astronomical measurements are where you learn if there's something wrong with our accounting of the amount of mass in galaxies, or our understanding of gravity.

Now, I'm not personally likely to discover any new physics, but I'm a physicist, and my bag is precision measurement. And it's a blast. It involves not only physics but also often optics, electronics, mechanics, math, and programming. It's recognized as a branch of physics, and is intensely experiment driven. Among the small handful of people who are crazy enough to be interested in all of those things at once, most are physicists.

Precision measurements are also an area where physics discoveries tend to escape the lab and end up in practical use. When a new effect is announced, I always wonder to myself: "When will they turn that effect into a sensor?" I can think of all kinds of measurements that benefit people, including GPS, atomic clocks, and blood oxygen sensors, that are practically all over the place.

I suspected I'd get an answer like this. Please read again: maybe there's not much low hanging fruit at the moment. Invictus0 has it right: this isn't the same thing as "no new discoveries ever".

Fields do have dry periods where for whatever reason, making progress is hard. Most famously AI had its "AI winter" for decades. We know now that there were new discoveries waiting to be made there but they required technology and datasets that simply didn't exist at the time. The last decade of AI progress has depended utterly on the growth of the public internet and then fast GPUs for processing that data. No matter how much funding the government had given symbolic AI in the 80s and 90s it'd have got nowhere. At least not on the use cases people seem to care about.

I find canned answers to this question increasingly tiresome. Academia and government funded research operates on a massive scale. It's unacceptable to me, as one of the people who actually pays for all this, that researchers entirely opt themselves out of any questions of utility or accountability. Fundamental physics in particular should take a good hard look at itself as it's both very expensive and in recent years, delivered very little. Consider string theory. It's been developed since the 1970s. 50 years now and for what? As far as I know this has delivered nothing concrete.

This isn't unique to physics, that's just an example. Epidemiology went down the statistical modelling rabbit hole 20 years ago and never emerged: I've yet to encounter a non-misleading claim coming from this field. If the entire field had been defunded 20 years ago we'd have a much healthier and saner world.

It's also not true that simply labelling something research means it'll one day be useful. Phrenology was once considered to be research. Critical race theory is labelled research. I think we can safely say these fields will never be useful and in fact have had sharply negative utility.

The 'normal' matter is 5% of the universe, do we need to know what the other 95% is? Do we need working nuclear fusion? Etc.

An incredibly short sighted view of academic utility. Maxwell's work on radio waves took 30 years to be developed into a "useful output".

The purpose of scientific advancement cannot be understood on the timetables of capitalistic utility.

Reply on news.ycombinator.com