SETI has decades of listening to its name, and detection of a intentional hello, but the most important revelation can be a message that the speaker is sending to you.
The recent effort to find extraterrestrial intelligences is paying more and more attention to “signal” as an umbrella term that encompasses any indicator of technology, not necessarily radio conversation. The present definition of technosignatures by NASA encompasses anything as intimate as radio or laser pulses all the way to the industrial waste materials of a nation’s industry and, most possibly, energy-harvesting megastructures that would broadcast themselves in the form of excess infrared “waste heat” (technosignatures). That growth is significant since it makes SETI an anomaly-hunting exercise: finding patterns that are inconsistent with the normal behavior of the sky, then demonstrating that they are not products of instruments, software, or anthropogenic interference.
“A glitch” in that respect is not a failure mode. It is the front door.
Radio SETI is already working in a regime in which human intuition is an inadequate filter. In a commonly known re-analysis of archival observations, deep-learned through 150 TB of data of 820 nearby stars, eight signals of interest were discovered by the algorithm that had previously been overlooked by classical methods (150 TB of data of 820 nearby stars). The events that the candidate had in common between themselves were: very narrow spectral width, non-zero drift rates indicative of relative acceleration, and behavior that occurred when the telescope was pointed at a target and disappeared when it was not. None of that amounts to confirmation, but it depicts a fundamental change in practice discovery starts not with a human seeing an overt beacon, but with computational triage.
This computer triage is more and more like the type of self-observation that the sophisticated devices employ to keep their machinery running, since it is upon these same pipelines that the spurious discoveries are produced. There are numerous causes of the anomalies produced by a radio observatory: wrong calibration, electronic malfunctions, weather, network hiccup, or an unusual astronomical event. It is not just a matter of finding oddities, but distinguishing between odd because the universe is fascinating and odd because the instrument is having a poor day. The work on system-health anomaly detection has started to resemble the SETI methodological twin. A single framework constructed on top of LOFAR spectrograms was developed to both categorize known failure modes and also indicate unknown rare behaviors, and was demonstrated to run at less than 1 ms per spectrogram in a real-time pipeline environment (less than 1ms to process a single spectrogram). Speed is not a luxury: it is what enables the inspection of “glitches” prior to the passing of the observing window and the sky shifting.
This reasoning is spreading beyond radio astronomy with the development of survey telescopes towards a data stream that is constantly on and community wide. The Legacy Survey of Space and Time of the Vera C. Rubin Observatory is configured in a fashion such that individual observations are not made by astronomers, but rather the project conducts an integral cadence and provides processed data products in a global manner ( All of the data obtained with Rubin Observatory ). That model is biased towards automated discovery: once billions of sources are repeatedly measured, what was rare, temporary or geometrically unusual will be observed primarily according to the algorithms that identify deviations earlier and justify deviations later.
By definition Technosignature searches reside on that edge between discovery and quality control. Minutes long spike; a suspiciously clean narrowband line; a light curve that is not like a planet each at first a glitch before it manages to pass the test of instrumentation. Practically the next SETI milestone can be not so much a deciphering of a cosmic postcard but a watch on an anomaly detector that declines to identify something as “normal” and then finds out that the decision is the outcome.
