June 14, 2026
GoPro gold or comment-section chaos?
I indexed 669 GB of my GoPro videos using my M1 Max computer and local ML models
He turned a mountain of bike videos into a searchable archive—and the comments got wild
TLDR: One creator used his M1 Max computer to sort hundreds of hours of bike footage into searchable highlights without sending files online. Commenters turned it into a mashup of praise, chip-performance arguing, and one extremely unserious question about “other” video libraries.
A cyclist with 2,207 GoPro clips and a serious rewatch problem decided to let his computer do the hard part: sort through nearly 669 GB of video on his own machine, find the exciting moments, and send the good bits straight into his editing app. On paper, it’s a nerdy DIY win. In the comments, though, it instantly became a mini soap opera about what this project really is, what people want it to do next, and of course, whether the whole thing could be used for something far less wholesome.
The strongest reaction? A mix of impressed, confused, and hilariously unserious. One commenter basically said this belonged on Hacker News Show HN, which is internet-speak for: cool build, wrong room. Another zeroed in on the machine itself, stunned that an Apple laptop-class chip could hang with a high-end Intel processor, then immediately dragged Windows into the debate. Classic comment-section behavior: a story about bike videos turns into a platform war in three moves.
And then came the quote everyone will remember: “Does it work for porn collections too?” Subtlety? Dead. The creator jumped in with a very earnest request for feedback on new features, while the crowd bounced between genuine curiosity and full gremlin mode. That’s the real vibe here: a clever local-video tool meets the internet’s unstoppable urge to turn every invention into a joke, a benchmark fight, or both.
Key Points
- •The article documents a local workflow for indexing a large personal archive of GoPro cycling videos.
- •The author says the full collection contains 2,207 GoPro videos that would otherwise require manual review.
- •The project runs on an Apple M1 Max computer using open-source machine-learning models locally.
- •The indexing system is designed to make footage searchable and identify interesting moments from long recordings.
- •The article reports indexing 628 videos totaling 668.68 GB and 15h 13m 18s of footage, with clips exportable to a DaVinci Resolve timeline.