I dislike watching videos (even at 2x). In fact, as soon as I got access to OpenAI’s GPT API access, I wrote a program that takes any YouTube video and generates 3 summaries of it (a 100-word summary, a 200-word summary, and a 400-word summary). I start with the shortest summary and if it is interesting then move on to the next one. (I rarely end up watching the whole video.)
I never wrote about it here because you’re not a programmer and you’re not going to be able to easily write and use that program.
FOMO no more. Google Bard (which is currently like a slightly dumber cousin of ChatGPT now natively does YouTube video translations.
See the full transcript here.
This wasn’t a contrived example I made up just for the purpose of this article. I was actually sent this video: normally I would have ignored it because tv;dw (something I’ve been doing for years), but now I used Bard to quickly see just find out what the last proof in this video was because that’s what I was told to look for.
Actually, come to think of it, I was being unnecessarily inefficient. Why do I need a summary of the full video. I could have simply asked for the last proof. Like so:
I want to keep reminding you is that the new GenAI technologies are very powerful and we are all underutilizing them.