1. WinRar’s profitability and business model vs. free alternatives
Commenters note that WinRar remains profitable despite being closed‑source, contrasting it with free tools like 7‑Zip.
- “They are a German GmbH and must publicly state their financials… Looks pretty profitable to me.” — shezi
- “winrar is profitable? sure? well, on the other hand, they sure don’t make losses.” — wolfi1
- “They're one of the few companies that actually manage to sell 'boxed software' (i.e. has not changed much in years but new customers keep buying it).” — amiga386
2. Technical merits of modern compression algorithms (zstd, gzip, lzma, etc.)
Many users praise zstd for its speed/compression ratio trade‑off and note its adoption in various projects, while also comparing it to gzip, lzma, and bzip2.
- “Everyone should be using zstd, and 7-zip does not support it.” — xxs
- “It's just the best general purpose compression algorithm, in terms of compression ratio to CPU used, for the vast majority of use cases.” — shawabawa3
- “Yeah, zstd is awesome. I built a webapp that uses it via wasm and the decompression speed is incredible…” — tnelsond4
- (Table showing lzma outperforming gzip/bz2 on repetitive data) — cgio
3. LLMs as compressors and the analogy between language modeling and compression
A substantial thread discusses whether large language models can be viewed as compressors, citing experiments where LLMs beat traditional compressors (e.g., xz) and debating the usefulness of such comparisons.
- “Quite well. This project[1], by Fabrice Bellard of ffmpeg fame, is quite old in AI years and uses an ancient LLM, but still beats xz by a solid margin.” — magicalhippo
- “If we must compare language models to compression, they are much more like jpeg and mp3 than they are gzip and flac.” — bob1029
- “Compression is a property of language… Every sentence has layers of compressed sentences.” — elendilm
- “I think language itself is compression… then definitionally, Language Modelling must also be Compression.” — adityaathalye