frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Frozen Function-agnostic goldenratio calculus (pip install phi-engine)

https://github.com/Purrplexia/LettersToMyHeroes/blob/main/Cantor_GoldenContinuum/code/phi_engine/README.md
1•purrplexia•38m ago
`pip install phi-engine`

Full Mathematical Proof:

https://github.com/Purrplexia/LettersToMyHeroes/blob/main/Ca...

Source Repo:

https://github.com/Purrplexia/LettersToMyHeroes

PyPI package:

https://pypi.org/project/phi-engine

Hi HN,

You are the first public audience to see this project. It has not been posted anywhere else yet. I chose to release it here first because Hacker News consistently rewards rigor, originality, and open-source tools built for ideas rather than for marketing.

This project contains entirely original mathematics. The core theory is proven in a rigorous 34-page proof titled "LetterToCantor.pdf" included in the repository and linked above. It builds a constructive continuum and a gridless analytic calculus from factorial moment laws and golden-ratio scheduling.

Derivatives and integrals of any analytic function can be computed *exactly* by a *frozen*, function-agnostic operator composed of exact rational coefficients (a "β-stream"). No grids, no step size, no limits, no symbolic math. Once synthesized, these operators work globally for every analytic function.

β-streams can be synthesized on-the-fly in milliseconds for thousands of digit accuracy, precompute is only needed if you want to push the boundaries of digits to comical scale.

Once a β-stream is synthesized or loaded, it is cached and reused on all future functions. No recompute, no wasted cycles, no dependency on the function.

To make this reproducible and verifiable, the project includes phi-certificates: canonical JSON/gz files that store these global calculus operators as exact rationals, sealed with canonical SHA-256. Anyone can load a certificate, verify its key material, and use the operator forever on any analytic function. The certificate is the calculus rule itself.

A phi-certificate is a canonical, global, arbitrary-precision analytic instruction, sealed by a 32-byte hash.

All code is GPLv3-or-later. All math documents are CC-BY-SA 4.0. Everything is permanently free and open.

The repo includes:

- the full mathematical proof (LetterToCantor.pdf)

- the phi-engine (exact rational contraction engine)

- phi-certificates (deterministic, hash-verifiable analytic operators)

- demos of extreme derivatives and integrals (sin(googol x^2) in 400ms to 6000 digits)

- demos of engine logic

- demos of certificate loading and emission

- full reproducibility scripts and PGP verification

Every artifact in this repo is hashed and PGP signed, ensuring bit-for-bit reproducibility.

If you want to see the math, start with the Letter (proof):

https://github.com/Purrplexia/LettersToMyHeroes/blob/main/Ca...

It is a crafted experience for those who love mathematics.

If you want to try the engine immediately:

pip install phi-engine

(see examples/ after install ive documented everything carefully for you)

If you care about reproducible, verifiable mathematics and globally reusable, cryptographically frozen analytic operators, I hope you enjoy exploring this.

I welcome questions, critique, and adversarial tests from the HN community — you are the first eyes on it — I will post it elsewhere later.

Verify it; test it; break it; disprove it.

Show HN: C++ for Autonomous Driving – From Learning to Landing AV Jobs

https://github.com/0voice/Awesome-CPP-Autonomous-Driving
1•ysy63874•48s ago•0 comments

Why Ceph and Rook Is the Gold Standard for Bare-Metal Kubernetes

https://oneuptime.com/blog/post/2025-12-03-ceph-rook-standard-bare-metal-storage-pools/view
1•ndhandala•5m ago•0 comments

How LLM Inference Works

https://arpitbhayani.me/blogs/how-llm-inference-works/
2•manishpushkar•5m ago•0 comments

Show HN: Kaomojihub – find your favorite kaomoji, learn meaning and usage tips

https://kaomojihub.net
1•adsl731898322•6m ago•0 comments

Linux on the Desktop

https://medium.com/@fluxusars/linux-on-the-desktop-eddde1973e29
3•fluxusars•7m ago•0 comments

Acme Challenge for Persistent DNS TXT Record Validation

https://datatracker.ietf.org/doc/html/draft-sheurich-acme-dns-persist-01
1•birdculture•9m ago•0 comments

I built an open source app to travel the world with AI

https://www.trytimetraveller.com/
1•rohitghumare•10m ago•0 comments

Waymo's Self-Driving Cars Are Suddenly Behaving Like New York Cabbies

https://www.wsj.com/lifestyle/cars/waymo-self-driving-cars-san-francisco-7868eb2b
1•Brajeshwar•10m ago•0 comments

The Data on Self-Driving Cars Is Clear. We Have to Change Course

https://www.nytimes.com/2025/12/02/opinion/self-driving-cars.html
2•mhb•12m ago•0 comments

Antithesis Raises $105M Series A Led by Jane Street

https://technews180.com/funding-news/antithesis-raises-105m-series-a-led-by-jane-street/
2•subvocalize•13m ago•0 comments

Minimal MCP Server Library

https://github.com/rcarmo/umcp
1•rcarmo•13m ago•0 comments

Remember When the Information Superhighway Was a Metaphor?

https://www.wsj.com/opinion/remember-when-the-information-superhighway-was-a-metaphor-bdd5a383
1•Bostonian•14m ago•1 comments

Frozen neural network shows intention signal in early layers (10× faster)

https://github.com/Anima-Core/an1-meaning-engine
1•ryananimacore•15m ago•2 comments

Coding a Snowflake Generator

https://cloudfour.com/thinks/coding-a-snowflake-generator/
1•paulhebert•16m ago•0 comments

How AI is transforming work at Anthropic

https://www.anthropic.com/research/how-ai-is-transforming-work-at-anthropic
2•mybestday•16m ago•0 comments

Win free access to a DGX B200 for 24h

https://lab.enverge.ai/b200-challenge/
2•tudorizer•16m ago•0 comments

W3C Web Neural Net API Draft

https://www.w3.org/TR/webnn/
1•bcardarella•16m ago•0 comments

Ask HN: Which post on HN has the most comments?

2•uneven9434•17m ago•0 comments

Explore Colors as You See: OKLCH in CSS

https://palettt.com/blog/oklch-in-css-why-is-it-better-than-rgb-and-hsl
1•mustafaiste•18m ago•0 comments

GOG Patrons

https://www.gog.com/en/patrons
5•nallerooth•18m ago•0 comments

Accelerating VMware migrations with a factory model approach

https://www.technologyreview.com/2025/12/03/1128488/accelerating-vmware-migrations-with-a-factory...
1•fleahunter•22m ago•0 comments

Helldivers 2 devs slash install size from 154GB to 23GB

https://www.tomshardware.com/video-games/pc-gaming/helldivers-2-install-size-slashed-from-154gb-t...
14•doener•23m ago•3 comments

Why the code agents should run inside the containers

https://envd.tensorchord.ai/blog/agent-container.html
2•urcyanide•23m ago•0 comments

DeepSeek-v3.2 Release

https://api-docs.deepseek.com/news/news251201
3•donpott•24m ago•0 comments

Joyboard is a balance board peripheral for the Atari 2600

https://en.wikipedia.org/wiki/Joyboard
1•doener•24m ago•0 comments

Show HN: Mapping DNS

https://loc.place
3•bo0tzz•28m ago•0 comments

A Devastating Fire Has Fueled a National Security Crackdown in Hong Kong

https://www.nytimes.com/2025/12/03/world/asia/hong-kong-fire-national-security-arrests.html
4•perihelions•31m ago•0 comments

YouTube Creators Find a New Consumer for AI Slop: Babies

https://www.bloomberg.com/news/articles/2025-12-03/ai-slop-youtube-videos-for-kids-pretend-to-be-...
3•thm•33m ago•1 comments

Artisanal – GitHub repos with no vibe coding activity

https://artisanal.vercel.app/
3•pipboyguy•34m ago•0 comments

Recommendations for Getting the Most Out of a Technical Book

https://sebastianraschka.com/blog/2025/reading-books.html
1•naves•35m ago•0 comments