Alien PostOffice

Learning to use Artificial Intelligence (AI) in the form of Large Language Models (LLM) has entirely changed my cost-curve. What before I might have wanted to explore, but would have taken more time than I wanted to spend, is now just a fun aside.

There is an old premise in science fiction that an alien race might gate first contact behind a proof of intelligence or capability. Arthur C. Clarke wrote stories where humans uncovered a probe buried on the Moon. Discovery of the probe meant humans had advanced enough to leave the Earth, and detect a buried probe on the Moon. This test does not change.

A more humorous story noted that (at the time) postal mail was faster across the country than local delivery. The character in the story addressed a letter to another star, which vanished instantly - and became the starting point for first contact. This test does also not change.

How would you uncover an alien postoffice?

In present, we might imagine an alien presence on the web, waiting to be discovered. What sort of clue or anomalous behavior might be a similar proof of intelligence?

This was the start of my IP_telescope experiment.

Imagine there are not-local folk who have attached to our internet. They might want to be found, given proof of intelligence. There are 2^32 possible combinations of IP addresses (or about 4 billion). There are local folk who (rightly) object to scanning the entire set.

How can we reduce this set? Well, prime numbers are a popular example.
(Not remotely assured correct, but a start.)

Choosing the usual four-part representation of an IP address, we can subset to addresses that are prime in each part. This reduces the set to about 7 million addresses. (Whether a not-local intelligence would choose this subset is unproven.)

(Defer to later - IPv4 addresses are a 32-bit number, and any prime inside 32-bits might be a beacon.)

If we probed this subset, might we find something unusual?

To be clear, we have not a clue as to what pattern to expect. So we have to look for things we do not expect. (So ... expect what you do not expect.)

The first theory is that an alien postoffice might be in an unassigned / disused space on the internet. (Given the noise made by the IPv6 folk about IPv4 space exhaustion, you might suspect - as did I - that this is not a thing. We were wrong.)

Wrote a tool (or asked to be written) to probe prime-part IP addresses. The tool would probe IP address that should not exist, but were live.

Found ... something.

To be clear, thinking about security requires a sort of induced but controlled paranoia. More than a few folk in that early community whose minds went a bit wobbly. They really are out to get you, except when they are not, and the clues are not always clear.

You have to be of more than one mind. On the one hand, you have the most-probable explanation mind. On the other hand, you have the how would I hide mind.

One story is that somehow an old website ended up abandoned, in a large space, and represents only a failed venture. This is a perfectly reasonable explanation.

Another story is that the website is a cover for something else. (Reasonable explanations are an advantage in this space.) If you wanted a scalable and obscured service, that fits.

Dunno. Not enough evidence.

Keep in mind, the bulk of the current CyberSecurity community is more about recipes (fixed mythology) and untroubled by theory. In this present, we are fixed firmly on common absense.

There is an off-chance that I might hear from Christian Gribnea, and he might prove believable. (Yeh. This is the more probable explanation. But not certain. Though ... a purple shirt ...)

And this is only the first attempt.