Show HN: I fixed FFmpeg's subtitle conversion (the bug from 2014)

FFmpeg—the free video Swiss‑army knife for media—just got a drama bomb. A solo dev dropped plug‑and‑play builds claiming they finally squashed a “2014 subtitle bug” #3819, added a Blu‑ray subtitle encoder (PGS), and unlocked 88 new conversions with 114 languages for text recognition. The demo flaunts karaoke‑style text that fades and moves, preserved when turned into Blu‑ray images. Sounds slick. But the comments? Spicy.

The top vibe is trust and receipts. One user fires the starting pistol with, “Where’s the source code?” and even posts a link to a GitHub branch. A veteran voice chimes in with the rulebook: if this heads to the official project, “disclose and describe the AI contribution” and add tests. Translation for non‑devs: if AI helped write the code, say so, and prove it works. Then the fact‑check hammer drops: “Open since 2014? No—filed and fixed in 2014,” calling out the headline’s timeline.

Meanwhile, the feature fans cheer the one‑command conversions, the Blu‑ray subtitle images with preserved effects, and the no‑dependency downloads built on FFmpeg 8.1. The dev touts continuous tests (FATE), plus two flavors: a minimal build and an English‑OCR build—with other languages dropped in via a folder. But skeptics side‑eye the binaries and nitpick the “fixed a 2014 bug” framing. Jokes fly—“where’s the sauce?” for source code—and a classic “2014 called, it wants its bug back.” The real plot twist isn’t the subtitles; it’s the clash between practical fixes and open‑source etiquette: source, AI transparency, and what counts as a “fix” versus a remix.