July 5, 2026
Pi-rates of the comment section
Connections in Math: the two kinds of random
Math blog says two random files can look alike — commenters instantly start a nerd civil war
TLDR: The post argues that two files can look equally random, yet one can still have a hidden shortcut if it was made by a simple rule like π. Commenters loved the brain teaser but instantly fought over whether the explanation cheats, whether π is even the right example, and whether the anti-AI disclaimer was doing a little too much.
A cozy math post about why a million digits of π can look just as random as a million die rolls somehow detonated into exactly the kind of comment-section chaos the internet was built for. The writer’s big idea is simple enough for non-math people: two giant files can seem equally pattern-free on the surface, but one might still be easier to recreate if it comes from a short recipe. In other words, one kind of “random” is about what the data looks like, and another is about whether there’s a hidden shortcut that made it.
But the real show was in the replies, where readers immediately began poking holes, nitpicking definitions, and lightly accusing the post of philosophical crimes. One camp said the post muddles what “entropy” can and can’t do, with contravariant arguing the real issue is that these measurements talk about a source of data, not one specific file. Another camp got existential fast: hyperhello complained that calling π “compressed” ignores the giant machine or human effort needed to actually generate it, basically asking whether the shortcut is only a shortcut because we’re pretending the rest of reality is free.
Then came the meta drama. andytratt deadpanned that the “no AI was used” disclaimer felt suspicious because the post had “many instances of claude,” which is exactly the kind of drive-by comment guaranteed to get screenshotted. And just when you thought the thread couldn’t get more delightfully nerdy, zzless reminded everyone that we don’t even fully know whether π’s digits behave as randomly as people love to claim. So yes: the math was deep, but the comments were deeper, pettier, and way more fun.
Key Points
- •The article compares a million-digit random file with the first million digits of π and says they are statistically indistinguishable by digit-frequency analysis.
- •It states that both files would pass randomness-style statistical tests despite differing in compressibility.
- •The post defines its discussion in terms of lossless compression, where the exact original data must be recoverable.
- •It distinguishes two forms of compressibility: statistical redundancy and short generative process descriptions.
- •The article introduces entropy as the measure for the statistical form of compression, based on average surprise in a symbol source.