4 ms·
The biggest takeaway for me was that it was getting info from YouTube videos. Is it actually watching and learning from the videos or where links to GitHub just
by 12907835202 4y ago
The biggest takeaway for me was that it was getting info from YouTube videos. Is it actually watching and learning from the videos or where links to GitHub just included in the comments?
- jdiff 4y agoTranscripts exist, and if there's multiple YouTube tutorials out there the odds are very good it also exists as a few dozen plain text articles.
- mgraczyk 4y agoI think this is just an incorrect assumption on the part of the blog authors.
- mtmail 4y agoCofounder here. We traced it back to two youtube videos where a developer is coding a phone tracking solution. The map then shows the geographical center of India and it's claimed that was the correct location. Then other users starting putting the code on github, then forked, then people created libraries and now AI tools pick that up. We already tried contacting the youtube authors, we left comments on the github repositories for months. Now we have to learn how to deal with ChatGPT. We also have no idea why a youtube author would describe a completely non-working solution over 20 minutes. Screenshots https://imgur.com/a/sNR87c7 https://imgur.com/a/sNR87c7 You can see the OpenCage logo on the bottom right of the images. We wrote a separate blog post about that about a year ago, we felt today's blog post would be too long if we added those screenshots, too.
- mgraczyk 4y agoBut these videos were not part of the ChatGPT training data. This is most likely caused by the code on GitHub. The claim in the blog post that "ChatGPT has picked up that content" is false.