Comments URL: https://news.ycombinator.com/item?id=47524451

Points: 3

# Comments: 2

https://suvakov.github.io/vibes/SlidingPuzzleChess/index.htm...

There are many ways to generalize them in 3D. Hope you’ll like my approach.

How it works:

The app visualizes a number N by recursively building a hierarchy of spheres based on its prime factors.

Given N with prime factors p_1, p_2, …, p_k: - Start with a single root sphere. - For the first factor p_1, place p_1 smaller spheres with their centers on the surface of the root. - For the next factor p_2, place p_2 even smaller spheres on the surface of each sphere from the previous level. - Repeat until all prime factors are used.

At the end, the number of smallest spheres is exactly N.

Sphere placement on each level is handled by a simple physics simulation using Coulomb-style repulsion, normalized by the size of spheres on the previous depth level, while keeping centers locked on the surface of the parent sphere. All spheres on the same depth level interact, not only those positioned on the same parent sphere. This way, spheres spread out evenly over the surface and in space. Sphere size is based on the minimum distance between spheres at the same level, scaled by an adjustable factor.

The code was written using Google Antigravity with the Gemini 3 Pro model.

Comments URL: https://news.ycombinator.com/item?id=46405305

Points: 3

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=43340295

Points: 2

# Comments: 0

      ___                        ___                     ___           ___     
     /  /\           ___        /  /\      ___          /  /\         /  /\    
    /  /::|         /__/\      /  /:/     /__/\        /  /::\       /  /::\   
   /  /:|:|         \__\:\    /  /:/      \  \:\      /  /:/\:\     /__/:/\:\  
  /  /:/|:|__       /  /::\  /  /:/        \__\:\    /  /:/  \:\   _\_ \:\ \:\ 
 /__/:/_|::::\   __/  /:/\/ /__/:/         /  /::\  /__/:/ \__\:\ /__/\ \:\ \:\
 \__\/  /~~/:/  /__/\/:/~~  \  \:\        /  /:/\:\ \  \:\ /  /:/ \  \:\ \:\_\/
       /  /:/   \  \::/      \  \:\      /  /:/__\/  \  \:\  /:/   \  \:\_\:\  
      /  /:/     \  \:\       \  \:\    /__/:/        \  \:\/:/     \  \:\/:/  
     /__/:/       \__\/        \  \:\   \__\/          \  \::/       \  \::/   
     \__\/                      \__\/                   \__\/         \__\/

https://news.ycombinator.com/item?id=42946317

You're absolutely right that some characters are missing even in those short books, and there are likely many more relationships that haven't been fully captured. That said, I’m still quite impressed by how much data the LLM extracted in a single pass, especially given the complexity of the task, the size of the input, and the strict output format.

My estimate of quality was subjective. To truly quantify accuracy, we’d need to establish a "ground truth" with a better approach and measure the difference between the generated and actual relationship graphs. One possible way to do that would be to process the text in multiple passes: first extracting characters, then identifying relationships, both steps with more sophisticated prompt engineering. Another way is to manually annotate the network. The only book I found with a publicly available, human-annotated character network is Les Misérables, based on Donald Knuth’s work: https://github.com/MADStudioNU/lesmiserables-character-netwo...

However, there is an additional challenge. Even with human annotation, the question remains: how to define relationship network? What is a relationship in a book? Should it be limited to explicitly stated connections in the text, or it also can include deduced relationships based on context with some probability? Defining these criteria is crucial to quantify quality of the result.

In this repository, you can find two tools: a script that extracts data from book text using an LLM (Gemini or OpenRouter API) and an HTML/JS (D3) visualization of the character graph. An external text-to-image model can be used to generate character illustrations (a Google Colab example is provided).

Explore the visualizations, play with the script, and feel free to add more books to the repo via pull requests!

Comments URL: https://news.ycombinator.com/item?id=42946317

Points: 9

# Comments: 4

The T2T team published a preprint [1] last December and released the data [2] in March. However, due to the peer review process, the findings have only just been formally published in Nature. The publication timeline can indeed be slow, and in cases like this one, the question is: what's the point when all scientists interested in the topic already know about it and working with this assembly?

[1] https://www.biorxiv.org/content/10.1101/2022.12.01.518724v1.... [2] https://github.com/marbl/CHM13

Only outlier is Djokovic's positive test. It perfectly fits to 26. December, the same date timestamp was created.

38 million km2 (Moon area) : 150 million km2 (land area of the Earth) = 25%