News for hackers. Stuff that matters.

lafalce

lafalce on Hacker News

Re:Triple-A Minesweeper on Friday October 09, 2026 @05:10PM (#50026591)
I need to sweep these mines!

Show HN: Cinematic Grade, a real-time GLSL color grade for the Cinnamon desktop

Posted by lafalce on Friday October 02, 2026 @05:30PM
from the show-hn dept.

https://github.com/mateolafalce/iron-within

Hey guys! I made a screen filter for the computer with the Iron aesthetic, because I really love what it conveys. I was already using my screen in black and white day to day, and I tweaked it a bit so it sits closer to the Iron look without draining the battery in the processgithub.com/mateolafalce/iron-withinI sent this applet to Linux Mint's official Cinnamon Spices repo, so maybe someday it...

( Read More... | no comments | on HN )
Re:Ask HN: What are you working on? (August 2026) on Sunday August 09, 2026 @02:15PM (#49233973)
I'm working on an open source project to make drone delivery accessible across any industry not only the software, but also the drone hardware
Re:Someone Build a OS for Claude Code on Monday June 22, 2026 @08:22PM (#48638476)
Basically โ€” to get AI-generated Hacker News comments with em dashes :)

Show HN: Recursive Language Model for Querying Human Action by Ludwig von Mises

Posted by lafalce on Friday January 16, 2026 @06:56AM
from the show-hn dept.

https://github.com/mateolafalce/human-action-rlm

Human Action by Ludwig von Mises is a 900+ page treatise on economics and praxeology dense, complex, and foundational to Austrian economics.I built this tool using Recursive Language Models (RLM) that ingests the entire book and lets you query it conversationally. It downloads the full text from mises.org, processes it with RLM for deep contextual understanding, and serves a simple Flask...

( Read More... | no comments | on HN )

Show HN: Turn any topic into a 3Blue1Brown-style video

Posted by lafalce on Friday January 09, 2026 @08:35PM
from the show-hn dept.

https://github.com/mateolafalce/topic2manim

Hey HN! I built Topic2Manim to automate the creation of educational videos like those from 3Blue1Brown.The workflow is simple:1. Give it any topic (e.g., "how ChatGPT works")2. An LLM generates an educational script divided into scenes3. LLM generates Manim code for each scene4. FFmpeg concatenates everything into a final videoCurrently working on TTS integration for narration!Would love...

( Read More... | no comments | 3 points | on HN )
Re:The Future of the Linux-rs Project on Sunday December 14, 2025 @04:57PM (#46267440)
After two years working on the linux-rs project, I've decided to pause development. I'm certain that not even thousands of developers could rewrite the Linux kernel from C to Rust within a decade.

With the LLM revolution, I believe the real challenge isn't manually rewriting the kernel, but researching and pursuing a recursive alternative through RL-based models.

Re:I'm Peter Roberts, immigration attorney who does work for YC and st... on Friday December 05, 2025 @02:21PM (#46166012)
Hi Peter, thanks for your time :)

I am planning to move to San Francisco to join the startup ecosystem. Regarding the H-1B visa, what does the current process and timeline look like for a new applicant? Given the lottery constraints, are there specific 'red flags' or common mistakes I should avoid when looking for a sponsor?

Re:My Contribution to Toon on Tuesday December 02, 2025 @09:45AM (#46121709)
As I mentioned on my X account, from my point of view, the TOON format shouldn't be utilized just like JSON (constructing data structures in .toon format to send to the LLM).

I believe TOON is a real game changer when used as an intermediary between the IDE/chat and the LLM. That hasn't been implemented yet, but it is undoubtedly the next step toward the project's mass adoption.

There will surely be news regarding this in the coming weeks :)

Re:Toon vs. JSON: Byte-Level Efficiency Model on Friday November 28, 2025 @07:11AM (#46077959)
Large Language Models increasingly rely on structured data for inference and function calling. However, standard formats like JSON introduce significant verbosity that inflates token usage and inference costs. This analysis presents a formal mathematical comparison between TOON and JSON to evaluate whether TOON achieves quantifiable efficiency gains by eliminating structural redundancy.

Under the assumptions described below (compact JSON, canonical TOON, ASCII keys and punctuation, shallow to moderate nesting, and mostly unquoted TOON strings), TOON's structural overhead is lower than compact JSON for the structure families analyzed here, except arrays of arrays.

Older ยป