What is btc-collision-hunter?

This is a small Python program that generates Bitcoin private keys one by one and checks whether each key corresponds to a wallet that holds any money. The project describes itself as "cryptographic performance art" and is upfront about the core mathematical reality: the chance of randomly hitting a funded wallet is so close to zero that you could run the program every second for trillions of times the age of the universe and still expect nothing. The whole point is that impossibility. Bitcoin's security depends on there being an astronomically large number of possible keys, around 2 to the power of 160 possible addresses. There are roughly 50 million funded addresses worldwide. Finding one by random guessing would require on average 2 to the power of 133 attempts. The program runs anyway, as a kind of collective proof of that limit. Anyone can download and run the script. By default it runs for one hour per day at a set time, checking 10 addresses per second. Participants report their running count to a shared GitHub issue, and an automated process aggregates the global total every 10 minutes. A public dashboard shows how many addresses the collective effort has checked so far, along with a running tally of how many wallets have been found (the expected answer: zero). On the privacy side, no private keys are ever saved to disk. Each key is derived on the fly from a counter value using a standard cryptographic function and thrown away immediately after the balance check. The counter itself is saved so the program can resume without repeating addresses it has already checked. This is not a practical tool for finding Bitcoin. It is an interactive art project built around demonstrating, through collective action, a mathematical fact that words alone cannot make feel real.