vrypan.eth

Down the RNG rabbit hole. https://blog.vrypan.net/2025/12/30/down-the-rng-rabbit-hole/

I built this. Still not sure how it compares to crypt/rand but it definitely looks comparable to math/rand and it's >400% faster. https://github.com/vrypan/rule30rnd

Result: 4x speed improvement, which makes the algorithm perform better than anything else available (to my knowledge). Proud human.

Can a totally predictable pattern show more entropy than a randomly generated one? https://blog.vrypan.net/2025/12/28/playing-with-turmites-better-than-crypto-rand/

Maybe everyone else knew about it, but I didn't: "parallel" is shell tool for executing jobs in parallel. It just saved me a couple of hours. https://www.gnu.org/software/parallel/

Two new turmite patterns added to the collection. https://art.vrypan.net/collection/turmite-patterns/

Any entropy experts here? I have generated an algorithmic, 100% reproducible image. It is 250x250 pixels, with 5 colors. The calculated Shannon entropy is 2.112490 bits, when 100% randomness would have something close to 2.321928 bits, i.e. 91% uncertainty. Is this something noteworthy?

Weekend digital cleanup. - Deleted AWS S3, Cloudfront, Route 53 configs I no longer used (I've moved many of my static sites to netlify+cloudfront) - Deleted GitHub repos used to host a single homebrew tap and moved most of them under https://github.com/vrypan/homebrew-tap. This repo also has a workflow to automatically update README(.)md when a tap is added or updated.

Most people you interact with don't use Direct Casts. Don't assume that sending them a DC will reach them.

Unusual book #2. This is a really weird one. The Voynich manuscript is an illustrated codex, hand-written in an unknown script referred to as Voynichese. The vellum on which it is written has been carbon-dated to the early 15th century (1404–1438). Stylistic analysis has indicated the manuscript may have been composed in Italy during the Italian Renaissance. The origins, authorship, and purpose of the manuscript are still debated, but currently scholars lack the translation(s) and context needed to either properly entertain or eliminate any of the possibilities. Hypotheses range from a script for a natural language or constructed language, an unreadable code, cypher, or other form of cryptography, or perhaps a hoax, reference work (i.e. folkloric index or compendium), glossolalia or work of fiction (e.g. science fantasy or mythopoeia, metafiction, speculative fiction). https://en.wikipedia.org/wiki/Voynich_manuscript

The Codex Borgia also known as Codex Borgianus, Manuscrit de Veletri and Codex Yohualli Ehecatl, is a pre-Columbian Middle American pictorial manuscript from Central Mexico featuring calendrical and ritual content, dating from the 16th century. It is considered to be among the most important sources for the study of Central Mexican gods, ritual, divination, calendar, religion and iconography. It is one of only a handful of pre-Columbian Mexican codices that were not destroyed during the conquest in the 16th century. Great gift for anyone interested in Mexican art. https://en.wikipedia.org/wiki/Codex_Borgia

This is perfect!

I wish projects used AI to build better documentation. Rn, 99.99% of the documentation out there is either: - written for script kiddies that just want to copy-paste code and hope it works. - (practically) internal technical documentation, written to be read as a reference for the people that are already deep in the project. - a mix of both, with 2 paragraphs explaining a complex technical concept as if readers have spent their life working on it, and a meaningless example in code that makes sense only if you use a very specific set of libraries, toolchain, and environment. Proper documentation used to be a difficult and boring process that required deep technical knowledge (and the devs that have this knowledge prefer to use it to write code, not documentation) and communication skills (and the people who know how to write comprehensive, easy to read documents often lack technical knowledge). It's easy to solve this now. A good technical writer can ask an AI to analyze the project code, and guide it to write good documentation. Sometimes, it's even better if the writer does not have expertise in the specific project, so they can understand what a newcomer needs.

The execution environment. https://blog.vrypan.net/2025/12/25/the-edges-of-llm-reality/

Cloudflare DynamicDNS can be really handy. And you get to use your own domains.

@boldleonidas wow! https://x.com/boldleonidas/status/2003355976830013561?s=46&t=xJuqzzzcZDWysPm0zVstxg

There are still things that make me optimistic. People that insist on the Cypherpunk values and ethos. by @pcaversaccio: https://x.com/pcaversaccio/status/2003124719403274279

@aave civil war. Good recap. https://x.com/obsrvgmi/status/2003089533995471074?s=20

"everything app" is the short version, for simplicity. It's actually "everything-generates-transaction-fees app". (Which is fine, this is their business.)

I like @unooo.eth so far! I have a feature request which is critical for me: better miniapp wallet handling. Ideally, I would like to connect my wallet during setup, and have Uno use it in a similar way Farcaster use warplet. I understand I'll have to switch apps for approvals, but I don't want miniapps to ask me which wallet, or think I don't have one.