home / github.com / item 47193832
Model Comparison
Model Editorial Structural Class Conf SETL Theme @cf/meta/llama-4-scout-17b-16e-instruct lite 0.00 ND Neutral 0.90 0.00 Technology @cf/meta/llama-3.3-70b-instruct-fp8-fast lite 0.00 ND Neutral 0.90 0.00 Tech Development claude-haiku-4-5-20251001 +0.19 +0.19 Mild positive 0.14 0.00 Knowledge Sharing & Collaboration
Section @cf/meta/llama-4-scout-17b-16e-instruct lite @cf/meta/llama-3.3-70b-instruct-fp8-fast lite claude-haiku-4-5-20251001 Preamble ND ND ND Article 1 ND ND ND Article 2 ND ND ND Article 3 ND ND ND Article 4 ND ND ND Article 5 ND ND ND Article 6 ND ND ND Article 7 ND ND ND Article 8 ND ND ND Article 9 ND ND ND Article 10 ND ND ND Article 11 ND ND ND Article 12 ND ND ND Article 13 ND ND ND Article 14 ND ND ND Article 15 ND ND ND Article 16 ND ND ND Article 17 ND ND 0.20 Article 18 ND ND ND Article 19 ND ND 0.20 Article 20 ND ND ND Article 21 ND ND ND Article 22 ND ND ND Article 23 ND ND 0.20 Article 24 ND ND ND Article 25 ND ND ND Article 26 ND ND 0.15 Article 27 ND ND 0.20 Article 28 ND ND 0.15 Article 29 ND ND 0.20 Article 30 ND ND ND
Summary Knowledge Sharing & Collaboration Acknowledges
This GitHub repository page documents SplatHash, an open-source image compression algorithm. The project engages tangentially with UDHR principles through MIT licensing (property rights), transparent technical documentation (education), and collaborative development structure (labor rights and community participation). The content is technically focused rather than explicitly rights-oriented, with mild positive signals toward intellectual freedom, knowledge sharing, and collaborative contribution.
Article Heatmap
Preamble: ND — Preamble Preamble: No Data — Preamble P Article 1: ND — Freedom, Equality, Brotherhood Article 1: No Data — Freedom, Equality, Brotherhood 1 Article 2: ND — Non-Discrimination Article 2: No Data — Non-Discrimination 2 Article 3: ND — Life, Liberty, Security Article 3: No Data — Life, Liberty, Security 3 Article 4: ND — No Slavery Article 4: No Data — No Slavery 4 Article 5: ND — No Torture Article 5: No Data — No Torture 5 Article 6: ND — Legal Personhood Article 6: No Data — Legal Personhood 6 Article 7: ND — Equality Before Law Article 7: No Data — Equality Before Law 7 Article 8: ND — Right to Remedy Article 8: No Data — Right to Remedy 8 Article 9: ND — No Arbitrary Detention Article 9: No Data — No Arbitrary Detention 9 Article 10: ND — Fair Hearing Article 10: No Data — Fair Hearing 10 Article 11: ND — Presumption of Innocence Article 11: No Data — Presumption of Innocence 11 Article 12: ND — Privacy Article 12: No Data — Privacy 12 Article 13: ND — Freedom of Movement Article 13: No Data — Freedom of Movement 13 Article 14: ND — Asylum Article 14: No Data — Asylum 14 Article 15: ND — Nationality Article 15: No Data — Nationality 15 Article 16: ND — Marriage & Family Article 16: No Data — Marriage & Family 16 Article 17: +0.20 — Property 17 Article 18: ND — Freedom of Thought Article 18: No Data — Freedom of Thought 18 Article 19: +0.20 — Freedom of Expression 19 Article 20: ND — Assembly & Association Article 20: No Data — Assembly & Association 20 Article 21: ND — Political Participation Article 21: No Data — Political Participation 21 Article 22: ND — Social Security Article 22: No Data — Social Security 22 Article 23: +0.20 — Work & Equal Pay 23 Article 24: ND — Rest & Leisure Article 24: No Data — Rest & Leisure 24 Article 25: ND — Standard of Living Article 25: No Data — Standard of Living 25 Article 26: +0.15 — Education 26 Article 27: +0.20 — Cultural Participation 27 Article 28: +0.15 — Social & International Order 28 Article 29: +0.20 — Duties to Community 29 Article 30: ND — No Destruction of Rights Article 30: No Data — No Destruction of Rights 30
Aggregates
Editorial Mean +0.19 Structural Mean +0.19 Weighted Mean +0.19 Unweighted Mean +0.19 Max +0.20 Article 17 Min +0.15 Article 26 Signal 7 No Data 24 Volatility 0.02 (Low) Negative 0 Channels E: 0.6 S: 0.4 SETL ℹ 0.00 Balanced FW Ratio ℹ 64% 25 facts · 14 inferences
Theme Radar
Foundation Security Legal Privacy & Movement Personal Expression Economic & Social Cultural Order & Duties Foundation: 0.00 (0 articles) Security: 0.00 (0 articles) Legal: 0.00 (0 articles) Privacy & Movement: 0.00 (0 articles) Personal: 0.20 (1 articles) Expression: 0.20 (1 articles) Economic & Social: 0.20 (1 articles) Cultural: 0.17 (2 articles) Order & Duties: 0.17 (2 articles)
HN Discussion
7 top-level · 7 replies
Thanks for sharing. I didn’t even know this type of thing had multiple algorithms.
Can you share what are the reasons someone may want to compress and image to 16 bytes?
Interesting, but my testing suggests that SplatHash is very weak at preserving global features, at least for synthetic images [1]. Both BlurHash and ThumbHash were able to preserve most of them, at the expense of worse (but still non-zero) local feature reproduction, but SplatHash simply discarded
all global features! I guess you need to store both local features (Gaussian splats) and global features (cosine bases) for the best result. The currently unused padding bit might be useful for that...
[1] I used my own avatars and icons as a test set. For example, https://avatars.githubusercontent.com/u/323836?s=400&v=4
so you turn images into colored bubbles? Why do people use this?
Very cool. To my eye, the splats are sometimes having too much contrast -- implying more "stark" visual features that don't actually manifest in the real image. Presumably the radius and the opacity curve of the gradients can be tuned to taste at the decoding phase, to make the splats softer?
The 6 blobs of colors look very weird after testing a few images, I feel like ThumbHash is much more natural and the downsides are minimal compare to SplatHash.
One other thing to compare it against is actual tiny JPEGs.
When you save a series of images as 16x16 JPEGs at the same JPEG quality level without optimization, you notice that there is a whole lot of common data between those files. Common data includes things like the file header (FF D8, FF E0 blocks), the Quantization tables, and the Huffman tables. If you cut away all the common data, the actual size of the image data is extremely tiny, usually under 64 bytes, though not a fixed size.
Here are the sizes of the four example images (just the unique image data) when resized to 16x16, then saved at quality 20:
First image: 48 bytes
Second image: 42 bytes
Third image: 31 bytes
Fourth image: 35 bytes
After appending back the 625 bytes of common data, you end up with a regular JPEG that can be decoded and displayed using fast native code from the browser.
ThumbHash page includes a comparison against "Potato WebP" which is probably a similar idea.
These things are called Low-Quality Image Placeholders (LQIP) and frequently used for front-end performance engineering.
For image placeholders while the real image is loading. At 16 bytes, that can easily be just another attribute on an html img tag.
definitely, will take a look
For privacy preservation and progressive revelation.
Inline it into a website's HTML to provide a low-res preview of an image as opposed to a blank placeholder or layout shift.
As a game engine dev, if I have an asset management app, it’s pretty reasonable that it might load the list of asset names and hashes before doing the significant work of decoding/generating thumbnails. This could give the app instant low quality thumbnails from loading the tiny array of data that’s already necessary just to get started.
Editorial Channel
What the content says
+0.20
Medium Framing
MIT license explicitly protects property and modification rights; frames intellectual property as open and collaboratively modifiable
FW Ratio: 60%
Observable Facts
Page displays 'MIT license' prominently in sidebar LICENSE file is listed and linked in repository structure CONTRIBUTING.md file exists in repository Inferences
MIT license ensures intellectual property remains protected while enabling collaborative modification Version control system respects contributor attribution for property rights +0.20
Medium Advocacy
Open-source project advocates for free expression through code sharing; README transparently documents technical ideas
FW Ratio: 60%
Observable Facts
Code is publicly available without access restrictions README explains project purpose and methodology ALGORITHM.md file contains algorithm documentation Inferences
Public code publication is a direct form of free expression and idea sharing Transparent technical documentation facilitates democratic sharing of ideas +0.20
Medium Advocacy Coverage
Collaborative development is highlighted; contributors are explicitly credited for their labor and intellectual work
FW Ratio: 60%
Observable Facts
Two contributors are listed with user links and names Repository shows 26 commits from multiple contributors CONTRIBUTING.md file exists in repository Inferences
Explicit attribution recognizes and respects workers' intellectual contributions Collaborative workflow structure demonstrates valuation of participant labor +0.20
Medium Framing Coverage
Project is presented as intellectual and cultural production; multi-language implementations demonstrate commitment to universal cultural access
FW Ratio: 67%
Observable Facts
Project provides implementations in three programming languages (Go, Python, TypeScript) Version history shows 26 commits from multiple authors Repository includes algorithm documentation and technical specifications Project is tagged with topics including image processing and related algorithms Inferences
Open-source model treats code as shared cultural heritage accessible to all Multi-language implementations demonstrate commitment to universal cultural participation +0.20
Medium Advocacy
Project structure emphasizes shared community responsibility and collaborative contribution; duties to community are embedded in open-source model
FW Ratio: 67%
Observable Facts
Multiple collaborators are listed as contributors Issues and Pull Requests sections are present on repository MIT license is applied to the project Benchmarks and documentation are included in repository Inferences
Project structure emphasizes shared community responsibility for maintenance and improvement License choice reflects commitment to community-oriented development standards +0.15
Medium Framing Advocacy
Educational content is prominent: algorithm explanation, technical documentation, installation guides for multiple languages, benchmarking methodology
FW Ratio: 67%
Observable Facts
README includes 'How It Works' section explaining the algorithm ALGORITHM.md file contains algorithm documentation Installation instructions provided for three languages (Go, TypeScript, Python) Benchmarks include comparison methodology and raw data Inferences
Structured technical documentation facilitates learning and knowledge transfer Multi-language implementations support educational access across different developer communities +0.15
Medium Framing
Open-source development model respects rights of participants to contribute and modify; collaborative norms are established through structure
FW Ratio: 67%
Observable Facts
CONTRIBUTING.md file exists in repository Multiple contributors are listed with equal attribution MIT license allows modifications and redistribution of code Issues section is visible on repository page Inferences
Transparent contribution process respects participant rights to shape development Open licensing framework respects collaborative modification rights ND
Page does not engage with human dignity, freedom, or justice frameworks
ND
No engagement with equal dignity or rights
ND
No discussion of discrimination or equality
ND
No engagement with life, liberty, or personal security
ND
ND
ND
ND
ND
ND
ND
ND
ND
No discussion of privacy, though image compression technology could have privacy applications
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Structural Channel
What the site does
+0.20
Medium Framing
GitHub structure enforces license-based property rights through version control and attribution tracking
+0.20
Medium Advocacy
GitHub's open repository structure enables unrestricted free expression through public code publication
+0.20
Medium Advocacy Coverage
GitHub's contributor attribution system respects and recognizes labor contributions through user linking and commit history
+0.20
Medium Framing Coverage
Open-source GitHub model treats code as shared cultural and intellectual heritage
+0.20
Medium Advocacy
GitHub architecture enforces community contribution norms through pull requests, issues, and transparent development
+0.15
Medium Framing Advocacy
Repository organization (docs/ folder, ALGORITHM.md, README) structures educational materials for public access
+0.15
Medium Framing
Contribution workflow and transparent licensing respect participant rights to modify and distribute code
ND
Repository structure does not address UDHR preamble themes
ND
No structural signals regarding human dignity
ND
No discriminatory barriers or practices observed
ND
No relevant structural signals
ND
ND
ND
ND
ND
ND
ND
ND
ND
No privacy-related structural signals
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Supplementary Signals
How this content communicates, beyond directional lean.
Learn more How well-sourced and evidence-based is this content?
0.87 low claims
Sources 0.8 Evidence 0.9 Uncertainty 0.8 Purpose 0.9
No manipulative rhetoric detected
0 techniques detected
Emotional character: positive/negative, intensity, authority
measured
Valence +0.3 Arousal 0.2 Dominance 0.6
Does the content identify its author and disclose interests?
0.50
✓ Author
More signals: context, framing & audience Does this content offer solutions or only describe problems?
0.91 solution oriented
Whose perspectives are represented in this content?
0.30 2 perspectives
Speaks: individuals
About: users developers
Is this content looking backward, at the present, or forward?
present medium term
What geographic area does this content cover?
global How accessible is this content to a general audience?
technical high jargon domain specific
Longitudinal
795 HN snapshots · 5 evals
Audit Trail
8 entries all eval pipeline all models llama-4-scout-wai llama-3.3-70b-wai claude-haiku-4-5-20251001
newest first
2026-02-28 16:04 eval_success Lite evaluated: Neutral (0.00) - - 2026-02-28 16:04
eval
Evaluated by llama-4-scout-wai : 0.00 (Neutral) 0.00 reasoning ED: Neutral tech presentation
2026-02-28 15:59 eval_success Lite evaluated: Neutral (0.00) - - 2026-02-28 15:59
eval
Evaluated by llama-4-scout-wai : 0.00 (Neutral) reasoning ED: Neutral tech presentation
2026-02-28 15:58 eval_success Lite evaluated: Neutral (0.00) - - 2026-02-28 15:58
eval
Evaluated by llama-3.3-70b-wai : 0.00 (Neutral) reasoning PR tech content
2026-02-28 13:22
eval
Evaluated by claude-haiku-4-5-20251001 : +0.19 (Mild positive) +0.02 2026-02-28 12:16
eval
Evaluated by claude-haiku-4-5-20251001 : +0.17 (Mild positive)