Avi Loeb is asking whether some UAP are leftovers from a technological civilization that lived on Earth long before us. I want that question kept alive, and I also want it argued properly, because the version circulating as a video clip skips the arithmetic that would make it useful.
The reasoning in the post is compact. Interstellar travel is punishing, so instead of asking a species on an exoplanet to survive a crossing that might take a large fraction of a billion years, you ask a machine to survive a few hundred million years in the neighbourhood of a planet we already know is habitable. Relics can outlive their makers, the same way our own autonomous systems would keep running for a while if we vanished. Orbit preserves hardware better than crust does. Then the win-win close: human-made means a national-security dividend, pre-human means we learn our own deep-time history, interstellar means we learn about the galaxy.
The idea has a serious pedigree, and it points somewhere else
Gavin Schmidt, who directs NASA GISS, and Adam Frank at Rochester published "The Silurian Hypothesis" in the International Journal of Astrobiology in 2019(arXiv preprint 2018). It is a detectability exercise, not a claim that anyone was here. Their finding is the part people skip: for anything older than roughly 4 million years, direct evidence(objects, fossils) is nearly hopeless, so the realistic search runs through chemistry, carbon isotope excursions, nitrogen from fertilizers, sedimentation and anoxia signals, with the PETM at ~55.5 Ma as the worked analogue. They wrote plainly that the hypothesis cannot be regarded as likely merely because no other valid idea presents itself. They also named it after a Doctor Who species, which tells you how seriously they wanted the label taken.
Jason Wright at Penn State went further in the same direction in 2018. If a prior technological species existed in the Solar System, Earth's surface is the worst place to look, because plate tectonics and erosion destroy the record. He argues for the Moon, for buried layers on Mars, for asteroids, places where an artifact could sit for 10⁸ to 10⁹ years. That recommendation cuts directly against the application Loeb is making. The literature says look where nothing moves. The video says look up at the sky over Virginia.
The preservation numbers are the actual argument
Loeb's orbit point is the one worth checking with a calculator, because he is right in direction and short by orders of magnitude in size. Vanguard 1, launched 1958, is the oldest human object still in orbit, and NASA's estimate has it re-entering around 2198, a ~240-year life. LAGEOS sits much higher and was built for longevity on purpose, carrying a plaque showing continental drift across its estimated ~8.4 million year lifetime. That is the outer bound of deliberate orbital persistence engineering that our species has managed, and it is one to two orders of magnitude below the 10⁸ years the pre-human scenario needs. Down on the surface it is worse. Oceanic crust is recycled by subduction and rarely exceeds ~200 Myr.
None of that kills the hypothesis. It does tell you the search has to be lunar or deep-sediment or isotopic, which is exactly where Schmidt, Frank and Wright already pointed. And there is a counter-argument Loeb's essay does not take on: an industrial predecessor would have drawn down the accessible fossil carbon we later found sitting there intact, and would have left a sharp isotopic spike. Schmidt and Frank note the flip side, that a long-lived civilization running on something other than fossil fuel might leave almost nothing. Fine. But then you are back to the chemical trail, not to objects in the air.
The Triassic-Jurassic line is where I part company
In the longer Medium version Loeb asks whether the end-Triassic extinction(~201.4 Ma), with its rapid carbon spike, 4~15 K of warming, roughly half of marine genera gone, was triggered by the technologies of an advanced civilization. That is the most checkable sentence in the piece and it does not survive the check. The Central Atlantic Magmatic Province was emplaced synchronously with the extinction, mercury anomalies across six geographically separated T-J sections indicate pulsed volcanism, and carbon-cycle modelling attributes the negative isotope excursions to mantle-derived plus thermogenic carbon. There is a mechanism, it has multi-proxy support, and it is published in PNAS. Adding an unobserved civilization on top explains nothing that the basalts do not already explain.
The win-win-win framing bothers me for a related reason. Every outcome is a success, which means the hypothesis makes no risky prediction. Compare it with what came in front of Loeb's own council on 7 July 2026: John W. Birks at CU Boulder has a preprint under review at Atmospheric Chemistry and Physics arguing that UAP orbs are magnetically confined dusty plasmoids produced by meteors, based on 508 NUFORC orb reports correlating with American Meteor Society fireball reports at better than 99% confidence. The council was reportedly unconvinced, on the grounds that dust clouds should not accelerate and debris tails should smear. That is what a real argument looks like. A model that can be wrong, and specific objections that could kill it.
The result his own telescope produced
Here is the fact I keep coming back to, and it went almost unreported. The Galileo Project's first observatory reconstructed about 500,000 aerial-object trajectories over five months. Around 16%, roughly 80,000, were flagged as outliers and reviewed by hand. 144 stayed ambiguous, and even those are described as probably mundane objects lacking range information. Zero confirmed anomalies. Metabunk noted the near-total silence that greeted the release. The speculation travels, the null result does not, and that asymmetry is a story about attention, not about physics.
The official record is not quiet either. AARO's Historical Record Report Volume 1(March 2024) reviewed US government UAP investigation back to 1945 and reported no case representing off-world technology, no evidence of reverse-engineering programs, no authentic UAP non-disclosure agreements. The FY2025 report, out around 23 July 2026, again found no evidence of exotic technology while leaving open cases, including a reported event with roughly 100 airborne objects and two likely uncrewed surface systems near Navy assets off Virginia. NASA's independent study team in September 2023 landed on the same wall: the binding constraint is data quality, sensor calibration, metadata, curation.
Where I actually stand
I do not treat those official findings as the end of the matter, and I would not want anyone to read this as a debunk. Agencies that spent decades classifying their own sightings do not get the benefit of the doubt from me when they announce there was nothing to classify, and the 19 February 2026 declassification directive plus the PURSUE releases from 8 May 2026 exist precisely because that record was not honest for a long time. I also think the deeper premise underneath Loeb's question is sound and undersold: our species is not obviously the first advanced thing this planet has carried, and the geological record is far too lossy to prove otherwise. That is not fringe, it is Schmidt and Frank's own conclusion about detectability, stated in a peer-reviewed journal.
What I do not accept is the shortcut from that premise to today's sightings, made on Medium rather than in a journal, wrapped in a framing where every outcome counts as a win. Loeb has a pattern with this. 'Oumuamua, the IM1 spherules recovered on a crypto-funded expedition and then shown by Desch and Jackson to sit on the terrestrial fractionation line, 3I/ATLAS which he rates a 4 on his own scale while conceding it is most likely a natural comet. Dramatic hypothesis, wide amplification, specialist rebuttal, partial retreat with the framing intact. The council he chairs is real and consequential, tasked in support of the transparency directive, reviewing only unclassified material, with no budget past travel reimbursement and Michael Shermer seated as the designated skeptic. It deserves better raw material than a re-run(he published nearly this same thesis in November 2023, then pegged to the end-Permian at 252 Ma).
As such, my ask is small and specific. Take the pre-human question to the places the preservation arithmetic actually allows: lunar regolith, buried Martian layers, sediment chemistry at the hyperthermals. Then let Schmidt, Frank and Wright say on the record whether they think their work supports the UAP application at all, because right now the only person making that link is the one with a QR code poster behind him. I would rather have one falsifiable claim about the Moon than three unfalsifiable ones about the sky.