Posted by A star that gets brighter is normal. A star that explodes is normal. A star that is sitting in a photographic plate from 1952 and is simply not there in a modern survey image is a problem, and the hunt for vanishing stars exists because problems like that occasionally turn into physics. Occasionally, much more speculatively, they get raised as places where extraterrestrial technology might show up in the data.
That second possibility deserves a careful hearing rather than either a headline or a sneer. There is a real scientific argument underneath it, made by a serious physicist, and it has been tested more than once.
Dyson’s Actual Argument
Freeman Dyson published a two-page paper in Science in 1960 called “Search for Artificial Stellar Sources of Infrared Radiation.” Almost everything popular culture did with it afterwards misses the point. Dyson was not proposing a solid shell around a star — he called that idea mechanically impossible in later interviews and was mildly irritated that it stuck.
His argument was thermodynamic. A civilisation whose energy demand keeps growing eventually runs out of planetary surface and starts collecting starlight directly, using a swarm of independent structures in orbit. If it does that on a large enough scale, the star gets optically fainter from outside, and that is the part everyone remembers. The part Dyson actually cared about was the consequence: the collectors absorb visible light, warm up, and must radiate the energy back out at longer wavelengths. So the search target is not darkness. It is an object producing far more infrared than its visible brightness accounts for.
Where the Energy Has to Go
This is what makes the whole idea testable rather than just entertaining. Energy is conserved whether or not aliens are involved. Intercept a star’s output and you have to dump the waste heat somewhere, and the temperature of the collectors determines the wavelength.
Which is also why the hypothesis is so hard to confirm. Warm dust does exactly the same thing. A star embedded in a debris disc, a young star still wrapped in its birth cloud, an evolved star throwing off shells of material — all of them look faint in the optical and bright in the mid-infrared. The natural explanation is common and the artificial one has never been observed, so the prior sits heavily on one side.
Nikolai Kardashev supplied the other half of the framework in 1964, sorting hypothetical civilisations by how much power they command: planetary, stellar, galactic. It is a crude ladder and Kardashev knew it, but it does one useful job. It tells you which observations could constrain the question. A civilisation using a planet’s worth of power leaves no astronomical trace at all. One using a star’s worth might. That is the entire reason anyone looks at stellar brightness for this.
The Case of Tabby’s Star
KIC 8462852 is the closest this field has come to a genuine public spectacle. Kepler watched it for four years, and volunteers on the Planet Hunters project flagged it because its light curve was bizarre — irregular dips, some as deep as twenty-two percent, with no periodicity and none of the clean symmetric shape a transiting planet produces. Tabetha Boyajian’s 2016 paper was titled, in effect, where is the flux going.
Jason Wright pointed out that a megastructure swarm would produce something along these lines, and the story went around the world. Then Boyajian’s team ran a crowdfunded observing campaign that caught the star dimming in real time across multiple filters, and found the dimming was wavelength-dependent — blue light blocked more than red. Solid objects block all colours equally. Fine dust does not.
So the answer was dust, most likely from a disrupted body. The episode is often written up as an embarrassment for SETI. It should not be. A hypothesis was proposed, a specific observation was designed to discriminate between it and the alternative, the observation was made, and the hypothesis lost. That is the process functioning exactly as intended.
Searches That Have Actually Been Run
The waste-heat idea has been surveyed systematically, which surprises people who assume this is all armchair speculation. Jason Wright’s G-HAT programme at Penn State went through mid-infrared data from the WISE satellite for roughly a hundred thousand galaxies, looking for entire galaxies whose starlight had been substantially reprocessed into heat. Nothing turned up that needed a technological explanation. That is a genuine constraint on how common galaxy-scale engineering can be.
More recently the Project Hephaistos team at Uppsala cross-matched Gaia, 2MASS and WISE catalogues across roughly five million sources and reported seven red dwarfs in 2024 with infrared excesses they could not immediately account for. Follow-up work has suggested at least some are blends with background galaxies. The candidates are not being announced as artificial by anybody involved — they are being published as objects that resisted the first round of ordinary explanations, which is the right way to hand something over.
Stars Do Vanish Naturally
Here is the counterweight, and it is a good one. In 2009 a red supergiant of about twenty-five solar masses in the galaxy NGC 6946 brightened modestly and then faded out of sight. Christopher Kochanek and Scott Adams, monitoring the galaxy with the Large Binocular Telescope, published it in 2017 as the best candidate yet for a failed supernova — a massive star collapsing straight into a black hole without the explosion, its light simply switching off.
JWST looked at the position in 2023 and complicated the story, with some analyses favouring a dust-shrouded stellar merger instead. Either way the object is genuinely gone from optical images, and either way the explanation is astrophysics doing something we did not fully understand.
Most vanishing stars will be this: variable stars caught at minimum, high proper-motion objects that moved, plate defects, or objects obscured by dust. A few might be black holes forming quietly. The base rate for exotic natural events is not zero, and it dwarfs the base rate for anything else.
The Pulsar Rule
Jocelyn Bell Burnell found a radio source in 1967 pulsing every 1.34 seconds with a regularity nobody could explain, and the Cambridge group labelled the folder LGM-1 — little green men — as a private joke while they worked out what it was. Rotating neutron stars, as it turned out. Nobody had thought of them.
The lesson taken from that episode inside SETI is procedural rather than deflating. An anomaly means the current explanations are inadequate, which is a statement about our models, not about the object. Extraterrestrial technology is one hypothesis on a long list and it carries the heaviest evidential burden of any of them, so it goes last, after variability, dust, instrumentation, astrometry and stellar evolution have all been worked through.
What would move a case forward is convergence: a source verified as real in several independent historical exposures, genuinely absent in deep modern imaging, with an infrared excess that does not fit a dust model, at a position with nothing else going on. Nothing has met that bar. The searching continues because the archives are large, the surveys are getting deeper, and the cost of looking is mostly computer time.
SETIworld follows the vanishing stars work as it develops, including the candidates that quietly turn out to be scratches on a photographic plate.
