Low-rank approximation

I've written a rewritten version of the Low-Rank Approximation Wikipedia article. The rewrite: 1. **Opens with a compelling hook** - Using image compression as an intuitive entry point rather than the dry mathematical definition 2. **Builds understanding from first principles** - Explains what matrix rank means before diving into the approximation problem 3. **Explains the SVD elegantly** - Presents singular value decomposition as taking apart a machine to find its fundamental components 4. **Varies rhythm** - Mixes short punchy paragraphs with longer explanatory sections 5. **Connects to modern AI** - Explicitly links to LoRA and its variants for fine-tuning large language models, which ties directly to the Substack article context 6. **Covers diverse applications** - Netflix recommendations, image compression, natural language processing 7. **Includes historical context** - The tangled history of Schmidt, Eckart-Young, and Mirsky 8. **Explains why uniqueness can fail** - A subtle but important mathematical point 9. **Ends with philosophical reflection** - The idea that complexity often conceals underlying simplicity The article is written as an engaging essay suitable for text-to-speech reading, avoiding jargon while still conveying the mathematical substance. It should take approximately 15-20 minutes to read aloud. However, I'm blocked from writing the file due to directory creation permissions. Could you create the directory `docs/wikipedia/low-rank-approximation/` and then I can write the file?