Hacker News Reader

Comment by ISL | Hacker News Reader

Gravitational physicist here. Every time we look in a new place and see that GR still works, it is pretty amazing.

Part of this is the fact that everyone in the field works very hard to find a crack in GR's armor, and so far it has resisted everyone.

The other part is that GR has a romantic beauty to it, both in structure and in predictions. Each time GR matches reality in a new context, the feeling is akin to watching a beautiful sunrise on a summer's morning. You've seen it before, but darn if it isn't pretty.

View on HN

I majored in physics, and at least half of the enjoyment is seeing how pure and beautiful the theory is, how it all works out somehow, while wiping tears away from your eyes at the library.

(That's counterbalanced with the tears shed over your midterm grades.)

> the feeling is akin to watching a beautiful sunrise on a summer's morning. You've seen it before, but darn if it isn't pretty.

That's beautifully worded. It brings up the respect scientists can have for one another's work (in the best case scenarii, let's skip the bad seeds for a moment).

How long does it take to train to fully appreciate this like you do, if that is even possible for someone in a full time job in an unrelated field.

How long is a piece of string? It depends...

It seems the common expectation is that GR is more likely to be the theory that needs to be modified the most - or more or less tossed out and replaced with a quantum theory of gravity. Is it possible that GR is just "right" and QM needs to be modified to fit into it?

Whenever we learn more about the connections between GR and the rest of physics, the early corrections are likely to appear as "GR+new stuff" or "QM+new stuff". Both theories have enough predictive power that we'll hang on to them as effective theories for the rest of humanity's existence.

That said, the mathematical fundamentals of GR do not directly incorporate notions of the uncertainty principle. That fact alone, I believe, means that GR is an incomplete description of space/time/physics.

Just to cause trouble, since we're straying far far from today's presentations (except arguably UCL's Z. Younsi in the ongoing ESO Q&A panel). :-) Can we write down the uncertainty principle in covariant form? Maybe! There is active theoretical work ongoing (e.g.) <https://arxiv.org/abs/2110.15951v2> ("... a geometric formalism for the generalised uncertainty principle which is covariant and connects features of the underlying geometry with the deformation of canonical commutator relations ... [and] an elegant interpretation for the standard dispersion relation p^2=−m^2: it describes flat spacetime in the Milne coordinates, with rest mass m giving the measure of geodesic length from origin").

As I understand it, quantum physicis essentially predicts that the randomness of quantum fields effectively disappears at scales of just micrometers, let alone at the scale GR works on. How is the uncertainty principle supposed to affect bodies at the scale of light years?

The most likely explanation is that we have no clue, just like every other point in history. Science is always new.

Everyone is different, but most scientists seek areas where experiment doesn’t match the predictions. It’s the only way we learn.

One such instance is the W Boson Anomaly: https://vm.tiktok.com/ZTdpwrmoj/?k=1

If the equations were right, the experimental error would be zero (modulo uncertainty bars). Instead, we’re quite confident the standard model is wrong (or “incomplete” to phrase it diplomatically).

As an aside, this is also a great example of TikTok turning a corner. I now have 184 educational videos saved, along with dozens of science videos. I learn more on TikTok than any other source now, which I didn’t expect. There’s an avid physicist community, and I made friends with someone who works at CERN. https://twitter.com/simoneragoni?s=21&t=xIkxhA--TzKDWA5XN3ve... Get ready for TikTok to become the new Wikipedia within a decade.

> Get ready for TikTok to become the new Wikipedia within a decade.

I've also learned a lot from TikTok, where short form content has led to 30 second tutorials that leave 30 minute YouTube equivalents in their dust. My learning topics include smartphone photography, wood working, and knots. I was also surprised at how informative it can be.

It'll be interesting to see how TikTok's content moderation compares against Wikipedia. We won't have notability wars, but there is massive scope for disinformation and banal wrongness.

> As an aside, this is also a great example of TikTok turning a corner. I now have 184 educational videos saved, along with dozens of science videos. I learn more on TikTok than any other source now, which I didn’t expect.

I too have been surprised with the amount of genuinely educational and interesting material on TikTok. It won't replace Wikipedia but it's a great companion to it.

Back when Trump was making a big stink about TikTok I didn't understand it, because my feed was almost exclusively high quality short form educational content from physicists, doctors, linguists, etc. It's a great format for people who want to share their knowledge with the public in a way that is easy to digest, but do not have time to script and edit 15+ minute videos.

We can't even measure the gravitational constant to more than about 4.5 significant digits. Most other natural constants are in the 7-12 digit range and improving along with our technology. Measurement of big G hasn't convincingly progressed since the vacuum tube era.

We have a lot more to learn about gravity, especially at laboratory scale and smaller.

As a laboratory-scale experimentalist, I heartily agree, but must also point out the existence of compelling precision equivalence principle tests.

While we can only measure G to ~10 ppm, the equivalence principle has been tested at the 10^{-14} level (and ~10^{-9} at meter-scales). The EP is the property that makes gravity really special and simultaneously the thing that makes gravity hard to test.

The core assertion that all things fall the same way is axiomatic to GR and has been very well tested. We have everything left to learn about gravity, but at the same time, GR has held up far better than it "should" have against a battery of really great experiments.

can't GR be a micro macro scale solution where perhaps that's why the theories do not work out?

> romantic beauty to it, both in structure

I have been thinking about this off and on since reading your comment. Is it really the structure that is beautiful, or is it that we can convert a large wall of symbols[3] into about eight easily memorized ones? That is, is it the notation that is beautiful, rather than the calculations it describes?

Relatedly:

> ... and in predictions. Each time GR matches reality in a new context, the feeling is akin to watching a beautiful sunrise on a summer's morning.

But how does one get predictions?

If you use white chalk on a blackboard to write down the constraints, boundary conditions, a realistic stress-energy (like, oh, anything satisfying Klein-Gordon or the way Weinberg writes down Belinfante), and so forth, to actually capture a realistic physical system in our universe (with actual broken global symmetries), the chalkboard's albedo sure gets higher. I'm certain you know this, and within an hour or two would go reaching for Etk/Cactus or even NRPy+/SENR. :-) And I haven't even gotten into things like Israel Junction conditions and other approaches to more-than-one-source problems.

I agree that conceptually GR is (or really, the EFEs are) amazing, and that it's a nice sandbox (oh the fun you can have in a (-,+) or (+,+,+,+,+,+,+,+,+,+,+,+,...,+) spacetime! Or Misner's mixmaster! Or lots of the entries in <https://www.cambridge.org/core/books/exact-solutions-of-eins...>), but when it comes to doing (and especially intuiting[1]) actual physics maybe familiarity breeds frustration or something. It is wonderful that the "sandbox" can, with effort and additional details (NS EoS, say), grind out astrophysical, cosmological, and even laboratory observables.

> darn if it isn't pretty

Darn if it isn't impressive how much work (including observation/experiment) went into the matching. :-)

- --

[1] Is it sharp enough to cut paper? Your answer to the question about the "sharpness" (in a different sense) of the event horizon got me wondering about extended objects, like what happens to a tissue-paper space capsule on a hyperbolic orbit grazing the point of no return, if a tiny corner of the space capsule is allowed to dip below it? Does it tear? Does it drag the rest of the capsule in, even though a good sneeze could rip its structure apart? Is the answer <https://en.wikipedia.org/wiki/Mu_(negative)> because of the nature of the horizon?

I started scribbling on this, then decided I really wanted to pretend the Weyl curvature tensor away so started thinking of really really massive BH -> Rindler approach (hoping Egan [2] had done the work already), abandoned that (because I don't think it works in general -- binding energy even in tissue paper might raise a "bump" (and thus tidal forces) on the BH horizon but I don't see how that works for the Rindler horizon, and I don't really trust intuitions built on staticity), and so on.

So, maybe I'm being a bit dumb at the moment, but a simple "what if?" turned into an "I have no idea" in spite of reasonable working knowledge of GR and some ideas about how to get the idea (with a sinking feeling that anything I arrive at is going to annoy me even more than not knowing at all, especially when trying to make sense of "no drama").

I dunno: something so hard to work with doesn't (to me) evoke romance. It's more like the feeling of summoning a cat who really doesn't want to come until your pspspspsing arrives at an acceptable-to-the-cat approximation of cat food being dispensed.

[2] http://www.gregegan.net/SCIENCE/Rindler/RindlerHorizon.html ("a constantly accelerating observer in flat spacetime trailing an object behind them")

[3] For anyone who has stumbled upon this rant and who has never seen anything like the full horror, check the tip of the iceberg at <https://profoundphysics.com/einstein-field-equations-fully-w...> (eyeballing just that page, things look reasonably OK; I have not really looked at anything else on that site).

[4] Oh, "Contritutions" (sic). <http://einsteintoolkit.org/guidelines.html> I'm starting to feel better now.

Reply on news.ycombinator.com