πfs: the data-free filesystem that “stores” data in π

πfs is an open-source filesystem project described as a “data-free filesystem.” Its core idea is that, since some mathematicians conjecture that pi, π, may be a normal number, every finite sequence of digits would theoretically appear somewhere within it, meaning any file could be treated as already existing at some position in π. The project README explains this in an extremely exaggerated, satirical tone: users no longer need to waste data on hard drives; they only need to remember a file’s position and length inside π, and they can “retrieve” the data when needed. In implementation, it follows a FUSE-style filesystem approach and requires tools such as autoconf, automake, and libfuse to build; using it also requires specifying a metadata directory and mountpoint. The key point is not that it provides practical compression, but that it deliberately pushes the boundary between “data” and “metadata” into absurdity: file contents are claimed to exist in π, but the system still needs to store position information for every byte or fragment, so the claim of saving data space is itself ironic. The README also notes that, to search for long sequences in practice, it splits files into smaller chunks, and in the current implementation even searches one byte at a time, which explains why it is extremely slow. The project also half-jokingly lists future directions such as variable-length search, arithmetic coding, parallel lookup, cloud-based π lookup, and πfs for Hadoop. For AI readers, its appearance in Hacker News AI keywords may simply be due to keyword matching or association with related projects; the original content does not propose a model, training method, or AI product. Its value lies mainly in technical humor, information-theory intuition, filesystem-interface experimentation, and reminding readers not to mistake conceptual “representability” for engineering “usability.”