I've been trying UBO lite and giving it a chance, but have been having nothing but weird issues with it. The internet doesn't work as well anymore, so thanks Google. I think it's time for me to mosey on over to Brave from Chrome finally.
what kind of nonsense are they doing with a link that uBO would block it? is it a 3rd party JS library that assembles an element that then places the link as stylized embed? I'd have expected more tracking type of stuff with it, but inspecting the element appears that the link is clean. my uBO did not block it????
>> More information: Shimpei Nishimoto et al, Infrared Bubble Recognition in the Milky Way and Beyond Using Deep Learning, Publications of the Astronomical Society of Japan (2025). DOI: 10.1093/pasj/psaf008
It links to a doi.org URL which directs the browser to what you linked.
> The DOI for a document remains fixed over the lifetime of the document, whereas its location and other metadata may change. Referring to an online document by its DOI should provide a more stable link than directly using its URL. But if its URL changes, the publisher must update the metadata for the DOI to maintain the link to the URL. It is the publisher's responsibility to update the DOI database. If they fail to do so, the DOI resolves to a dead link, leaving the DOI useless.
I’ve met the folks behind DOI. Very nice people (Jonathan Clark in particular).
It’s an independent foundation and they have backups/contingency plans established with major universities to preserve the DOI records in the event the foundation fails.
DOIs exist so they can be human readable and simultaneously indicate the source and veracity of it. They’re somewhat gated as well which serves a function.
* An auto increment ID is just as human non-readable as a UUID, it's just easier to get silent collisions from typos.
* The Source is metadata that belongs in a metadata system, not into the ID itself
* the veracity is worthless without verifiability
* gated-ness is just an anti-feature caused by the lack of verifiability
If you you classify identifiers along different axis of their properties, you'll notice that DOIs actually inhabit the completely wrong quadrant for their use-case.
(https://docs.rs/tribles/0.5.1/tribles/id/index.html)
Are they human readable? As for veracity, wouldn't baking a digital signature into the paper itself be far more reliable?
Don't get me wrong, I'm glad they exist, but they appear to guard against humans who are lazy and make mistakes sometimes rather than against a powerful adversary motivated to interfere with science. It might be time for an upgrade.
saint_yossarian 2 hours ago | root | parent | next [–]
It's blocked by "EasyList – Newsletter Notices", the annoyances lists are often a bit too aggressive.
I truly hope that the common theme of the likes of "JWST Just Found Something Which Should Not Exist" etc will not be augmented by stuff like "we used AI(tm) to figure out X, Y, Z".
The last thing we need is hallucinations fucking up the more grounded astrophysics. I'm not saying that is what is happening, I just worry about stuff like this. AI causing us to bark up the wrong tree, and so forth.
If anything, it's just going to call out a thing in the image that humans can then go and look at. Nothing in astronomy is ever "decided" by a single report. It gets looked at and scrutinized, and then committee style decisions are made about it. So if someone is using some ML to scan every image taken by JWST and calls out 1 cool thing for every other 9 things it finds that's "yeah, we know about that", then that's still quite a lot of new cool things. it'll just be able to do this faster and potentially much more in-depth than a human scanning across the images manually
Yeah but what if we start seeing only using this new awesome tool? What if that becomes the new seeing apparati? THIS is the tool that breaks that mold? The tool that (near?)every field is also going to be considering to be the tool that's off limits, or be 'constrained?'.
What if we had that view with microscopes, back when?
I see the point being made above fully. If ai takes over it's because we are every day it seems like slowly placing that faith.
It's our wow. It's the future generations taken for granted.
That's going to be a sign of the times if that happens. There are way too many people that enjoy the search doing it by hand. Yes, they are all of a certain age. Those of a certain younger age that only knows digital tools and not the ways of using their own eyes might eventually happen, but thankfully I won't be around for that to happen. (I'm one that uses my own eyes).
I can't tell if you're agreeing with my 'poibt' or disagreeing.
But yes, like telescopes. Or microscopes. Those still bind us to using our built in sensors that we 'trust'.
Then we obviously get into radio telescopes, or down to electron microscopes, etc and we start having to believe in the tech to get our new found understandings.
My mental hesitation lay in trusting AI to get to that level of belief -- if/when that happens, what do we really know or trust?
We've been using electron scanning microscopes, radio telescopes, etc for much much longer than we've been using "AI" in them.
I'm really not sure what you're getting at here, but you definitely seem to be confusing generative AI here. What's being discussed here is not generative AI. It's just a very refined algo searching for patterns in images. This is not "artist conception" type of content like the image of the black hole. So until you accept the difference, you're just spinning your wheels
> The last thing we need is hallucinations fucking up the more grounded astrophysics.
You're thinking of the wrong ML. Generative models "hallucinate" and it's as much a feature as it's a bug. ML in astrophysics is not generative. They use it for flagging, "binning" data and in general (simplified) classification.
Machine learning (AI) is used everywhere in astronomy. That's how they made the black hole image. Don't confuse the broader 60+ year old world of ML with transformers and diffusion models.
Based on the paper I linked, it seems like a straight up classical sampling and clustering with baysian hyperparameter tuning. This is “everything is now AI” slop that’s infected all grants, academic and private industry fundraising. There’s no neural net or LLM involved.
Clustering alone is machine learning and has been taught as such to innumerable people.
I have deep feelings about this, someone in management taking exactly one Kaggle course managed to wield this knowledge to great damage.
But it is machine learning.
Additionally, it goes far beyond clustering: the article you linked describes training an image recognition model, which also seems to be heavily stressed in the article linked on HN.
Eventually youll give in to the fact that ai is useful, and maybe revolutionary. Until then, continue using swear words and sticking your head in the sand
Yeah. Thanks for saying this. Please let be the real sciences real that have propelled the humanity forward with painstakingly detailed analysis by peer reviews and what not.
Let's keep AI for vibe coding, cat images and memes etc.
'Bubbles' could imply something like space-time changes, and in face a couple of comments speculatively / hopefully read them like this: [1], [2]
But they're (sadly?) much simpler: Spitzer bubbles 'are formed by radiation and winds from massive stars, which carve out holes within surrounding dust clouds.'
I would _love_ it if either of both these comments were true, by the way. Space-time can be boring and restrictive. What if...? I love the idea of bubbles reflecting a smaller universe and what it might hint about FTL, for example, and I live in hope that we'll find abberations and abnormalities like this.
Bubbles of expanding change in the structure of spacetime. Spoiler ahead. In the book, under the "Dark Forest" game theory setting, super advanced civilizations wipe lesser ones out in increasingly complex ways, for example, triggering a spacetime flattening in the target solar system, creating a bubble where s-t collapses into 2D. Also, scars in s-t when reckless young civilizations use FTL technology.
You wanna hear my evidence-free cosmic structure theory? Of course you do.
If you shine a laser through a mass of soap bubbles it will unsurprisingly split into lots of smaller beams due to a mix of refraction and reflection. I have long held the suspicion that there's an isomorphism between gravitational and surface tension structures, that the multiplicity and distance of galaxies may be somewhat illusory, and that many of them are translated/rotated reflections of nearer ones. Laugh now, perhaps gasp in wonder later.
There was a somewhat similar search for these duplicate galaxies as evidence for a universe with positive curvature. Because in that case if you look deep enough you'll see more images of the same galaxies although they'll be further back in time and possibly shifted in the way you're describing by the cosmic structure. It didn't pan out obviously.
I don't laugh but it is an interesting idea. Most of the theoretical physics starts that way and then gradually verifying such assumptions with great care and experimentation over multiple generations of scientists.
This concept is a bit too advanced for me (or the page is too minimal to easily understand), but it sounds fascinating. I'll read up more on it, thanks.
Fair critique! I tried to find a more accessible Wikipedia article but they all look like this...
Simply put, it's a topological defect or discontinuity, but that makes it sounds worse than it really is. I find it easiest to visualize with magnets. They want to align with their neighbors, so in general you get big blocks (domains) where all particles are aligned. What happens at the border when blocks with different alignments meet? We call that a domain wall. That's literally all it is!
You can find domains walls in magnets, metallic crystal grain structures, liquid crystals, pretty much anything that wants to self-align. One issue: gravity doesn't particularly want to self-align. Or does it?
What you're describing sounds like the curvature or topology of space would be non-flat. AFAIK this hasn't been completely ruled out, but so far every piece of evidence suggests the universe is flat over vast distances.
Intuitively I'd say if there was curvature or topological irregularities at the furthest distances we can observe, there wouldn't be a consistent redshift observed on far objects because some of them would be coming towards us instead of pulling away.
What you're describing is gravitational lensing. It can make one galaxy appear to be several in different places or shapes. It is, however, well understood.
I know what gravitational lensing is, but that's not what I have in mind (or rather, my gut - while I have a strong hunch about this, I do not want to invest the years of hard study to validate it or more likely end up in a dead end).
My hunch is that rather than space being a contiguous void with isolated mass of gravity behaving like tiny monopolar magnets, at the intersection between different mass systems there are 'surfaces' of some sort like the walls of a bubble in a pile of foam, and that if you could encounter this 'surface' you would either be repelled by it (most likely) or make contact and be able to slide around on it, and then once you got to the angles where walls joins, you would be able to zip along the intersections at great speed in ways that defy conventional physics. I can't really explain it in greater depth, it's an intuition that's half lifelong fascination with looking at soap films and what foam does, and half 'it came to me in a dream.'
Your comment reminds me of a picture I saw a few days ago of a telescope shot, caption "there are no stars in this picture, only galaxies" and there were so. Many. Bright spots.
I don't know where or when it was taken, or what part of the sky that happens in. Maybe it's just a really long lens, so it's seeing "through" the galaxy we normally see "stars" from?
Anyhow, how do you think you could prove this or how someone could prove it? Is it like, two observers on opposite sides of the planet observing the same thing, say during an eclipse or something? Maybe radioastronomy?
You might be thinking of the Hubble Ultra Deep Field image (https://science.nasa.gov/asset/hubble/hubble-ultra-deep-fiel...). I believe it was the result of NASA saying "Let's look at what appears to be a completely dark spot in space, zoom in as far as we can, and see what's there."
We see stars in our galaxy because they are close enough to us that we see them as individual stars. Compare that to the Andromeda Galaxy, which is far enough away that without intense zoom, it looks like a single source of light. There are galaxies even farther away, which we cannot see with the naked eye at all, but zooming in on them like Hubble did means we eventually get enough resolution to see they are individual galaxies, unfathomably far away.
JWST being able to see infrared means we'll see galaxies that are so far away, their light is redshifted so we (and Hubble) cannot even see them at all.
With regards to your question about how to test the bubble hypothesis posted by parent, we would be limited by how variable our point of view can be. We can gather what data we can at one end of Earth's orbit, and then try to see from the opposite end and compare what we see, comparing data sets to see if certain galaxies or stars are in different positions. We already do some of this when dealing with gravitational lensing and I believe it's one of the primary ways we can detect black holes, as they bend light a lot.
I really don't know how you'd prove it, or I'd be operating a cranky social media account demanding the scientific community pay attention to me! My knowledge of astronomy/cosmology tops out around Quanta magazine reader level and studying the subject academically always seemed like a luxury for people who already have money. The 'space is foam' idea just hit me out of the blue one day ~20 years ago when I was staring at bubbles and looking at how tiny ones interact with larger ones. I feel like it has something to do with heliopause and plasma, but every time I read up on it the scientific consensus seems to be 'we still don't know much about it,' so I don't know what to do with the idea.
> once you got to the angles where walls joins, you would be able to zip along the intersections at great speed in ways that defy conventional physics.
There's the assumption, at least by me in the past, that every image ever taken by any telescope has been poured over that nothing new could be found by someone like me looking at it. It wasn't until I realized that most images are looked at by the people capturing the image while they look at the image for the one thing they were trying to study. In a Hubble/JWST type image, that point of interest might be < 10% of the captured data. (Think of all of those images of new discoveries that have been so zoomed in that it's nothing but a bunch of pixels) Once they finish with it, it just becomes part of the archive. There have been lots of discoveries of people combing over the archives to find things in existing data without ever needing any 'scope time of their own.
> In 2010, Penrose and Vahe Gurzadyan published a preprint of a paper claiming that observations of the cosmic microwave background (CMB) made by the Wilkinson Microwave Anisotropy Probe (WMAP) and the BOOMERanG experiment contained an excess of concentric circles compared to simulations based on the standard Lambda-CDM model of cosmology, quoting a 6-sigma significance of the result.
These are probably mostly supernova remnants, but also, if you believe the "inside" (i.e. other side) of Black Holes are White Holes, it would make perfect sense to see White Holes in our universe where stuff is sort of "falling into" our universe and exploding outward. There might be an actual hierarchy of these kinds of things so that what we think was our "Big Bang" was actually the location where a White Hole emerged thru which flowed everything in our universe.
It's blocked by "EasyList – Newsletter Notices", the annoyances lists are often a bit too aggressive."
It links to a doi.org URL which directs the browser to what you linked.
> The DOI for a document remains fixed over the lifetime of the document, whereas its location and other metadata may change. Referring to an online document by its DOI should provide a more stable link than directly using its URL. But if its URL changes, the publisher must update the metadata for the DOI to maintain the link to the URL. It is the publisher's responsibility to update the DOI database. If they fail to do so, the DOI resolves to a dead link, leaving the DOI useless.
More about it at Digital Object Identifier (DOI) Under the Context of Research Data Librarianship - https://doi.org/10.7191%2Fjeslib.2021.1180
which reminds me, who has control over DOI.org ... eg. is it DOGE-safe ? likewise arXiv .. can it easily be co-opted / subsumed ?
It’s an independent foundation and they have backups/contingency plans established with major universities to preserve the DOI records in the event the foundation fails.
https://www.doi.org/the-foundation/board-and-governance/
(Third foundation?)
DOI must die
But magnet links and the BitTorrent mainline hash-table are a better DOI than DOI.
* An auto increment ID is just as human non-readable as a UUID, it's just easier to get silent collisions from typos.
* The Source is metadata that belongs in a metadata system, not into the ID itself
* the veracity is worthless without verifiability
* gated-ness is just an anti-feature caused by the lack of verifiability
If you you classify identifiers along different axis of their properties, you'll notice that DOIs actually inhabit the completely wrong quadrant for their use-case. (https://docs.rs/tribles/0.5.1/tribles/id/index.html)
Don't get me wrong, I'm glad they exist, but they appear to guard against humans who are lazy and make mistakes sometimes rather than against a powerful adversary motivated to interfere with science. It might be time for an upgrade.
detect-ability of state-actor post-facto editing : DEI related or otherwise
saint_yossarian 2 hours ago | root | parent | next [–] It's blocked by "EasyList – Newsletter Notices", the annoyances lists are often a bit too aggressive.
The last thing we need is hallucinations fucking up the more grounded astrophysics. I'm not saying that is what is happening, I just worry about stuff like this. AI causing us to bark up the wrong tree, and so forth.
What if we had that view with microscopes, back when?
I see the point being made above fully. If ai takes over it's because we are every day it seems like slowly placing that faith.
It's our wow. It's the future generations taken for granted.
"Much more in-depth" ways now just "the way".
Like telescopes?
But yes, like telescopes. Or microscopes. Those still bind us to using our built in sensors that we 'trust'.
Then we obviously get into radio telescopes, or down to electron microscopes, etc and we start having to believe in the tech to get our new found understandings.
My mental hesitation lay in trusting AI to get to that level of belief -- if/when that happens, what do we really know or trust?
I'm really not sure what you're getting at here, but you definitely seem to be confusing generative AI here. What's being discussed here is not generative AI. It's just a very refined algo searching for patterns in images. This is not "artist conception" type of content like the image of the black hole. So until you accept the difference, you're just spinning your wheels
You're thinking of the wrong ML. Generative models "hallucinate" and it's as much a feature as it's a bug. ML in astrophysics is not generative. They use it for flagging, "binning" data and in general (simplified) classification.
For another ML-assisted science thing, there's the LHC: https://en.wikipedia.org/wiki/Higgs_boson#Findings_since_201...
Clustering alone is machine learning and has been taught as such to innumerable people.
I have deep feelings about this, someone in management taking exactly one Kaggle course managed to wield this knowledge to great damage.
But it is machine learning.
Additionally, it goes far beyond clustering: the article you linked describes training an image recognition model, which also seems to be heavily stressed in the article linked on HN.
Let's keep AI for vibe coding, cat images and memes etc.
But they're (sadly?) much simpler: Spitzer bubbles 'are formed by radiation and winds from massive stars, which carve out holes within surrounding dust clouds.'
So really just the blast radius!
-- https://www.spitzer.caltech.edu/image/ssc2013-05a1-bubbles-w...
I would _love_ it if either of both these comments were true, by the way. Space-time can be boring and restrictive. What if...? I love the idea of bubbles reflecting a smaller universe and what it might hint about FTL, for example, and I live in hope that we'll find abberations and abnormalities like this.
[1] https://news.ycombinator.com/item?id=43552920
[2] https://news.ycombinator.com/item?id=43552713
If you shine a laser through a mass of soap bubbles it will unsurprisingly split into lots of smaller beams due to a mix of refraction and reflection. I have long held the suspicion that there's an isomorphism between gravitational and surface tension structures, that the multiplicity and distance of galaxies may be somewhat illusory, and that many of them are translated/rotated reflections of nearer ones. Laugh now, perhaps gasp in wonder later.
Sounds like domain walls.
https://en.wikipedia.org/wiki/Domain_wall
Simply put, it's a topological defect or discontinuity, but that makes it sounds worse than it really is. I find it easiest to visualize with magnets. They want to align with their neighbors, so in general you get big blocks (domains) where all particles are aligned. What happens at the border when blocks with different alignments meet? We call that a domain wall. That's literally all it is!
https://en.wikipedia.org/wiki/Topological_defect
https://en.wikipedia.org/wiki/Magnetic_domain
https://en.wikipedia.org/wiki/Domain_wall_(magnetism)
You can find domains walls in magnets, metallic crystal grain structures, liquid crystals, pretty much anything that wants to self-align. One issue: gravity doesn't particularly want to self-align. Or does it?
https://en.wikipedia.org/wiki/Higgs_mechanism
https://en.wikipedia.org/wiki/Spontaneous_symmetry_breaking
I've only got a surface understanding of this stuff myself. Best of luck in your research!
Intuitively I'd say if there was curvature or topological irregularities at the furthest distances we can observe, there wouldn't be a consistent redshift observed on far objects because some of them would be coming towards us instead of pulling away.
My hunch is that rather than space being a contiguous void with isolated mass of gravity behaving like tiny monopolar magnets, at the intersection between different mass systems there are 'surfaces' of some sort like the walls of a bubble in a pile of foam, and that if you could encounter this 'surface' you would either be repelled by it (most likely) or make contact and be able to slide around on it, and then once you got to the angles where walls joins, you would be able to zip along the intersections at great speed in ways that defy conventional physics. I can't really explain it in greater depth, it's an intuition that's half lifelong fascination with looking at soap films and what foam does, and half 'it came to me in a dream.'
I don't know where or when it was taken, or what part of the sky that happens in. Maybe it's just a really long lens, so it's seeing "through" the galaxy we normally see "stars" from?
Anyhow, how do you think you could prove this or how someone could prove it? Is it like, two observers on opposite sides of the planet observing the same thing, say during an eclipse or something? Maybe radioastronomy?
We see stars in our galaxy because they are close enough to us that we see them as individual stars. Compare that to the Andromeda Galaxy, which is far enough away that without intense zoom, it looks like a single source of light. There are galaxies even farther away, which we cannot see with the naked eye at all, but zooming in on them like Hubble did means we eventually get enough resolution to see they are individual galaxies, unfathomably far away.
JWST being able to see infrared means we'll see galaxies that are so far away, their light is redshifted so we (and Hubble) cannot even see them at all.
With regards to your question about how to test the bubble hypothesis posted by parent, we would be limited by how variable our point of view can be. We can gather what data we can at one end of Earth's orbit, and then try to see from the opposite end and compare what we see, comparing data sets to see if certain galaxies or stars are in different positions. We already do some of this when dealing with gravitational lensing and I believe it's one of the primary ways we can detect black holes, as they bend light a lot.
this is what i was thinking of; thanks for the hubble reference, though!
Hyperspace lanes!
https://x.com/jharohit/status/1479100020049678339?s=46
Great use of AI!
> In 2010, Penrose and Vahe Gurzadyan published a preprint of a paper claiming that observations of the cosmic microwave background (CMB) made by the Wilkinson Microwave Anisotropy Probe (WMAP) and the BOOMERanG experiment contained an excess of concentric circles compared to simulations based on the standard Lambda-CDM model of cosmology, quoting a 6-sigma significance of the result.
https://en.wikipedia.org/wiki/Conformal_cyclic_cosmology