Nice to see this here. I’m the developer who made this. The idea for this started when I was chatting with co-creator, Dave Charney and we asked, “I wonder if all the questions kids could ask of astronauts have already been asked.”
If you select “About this project”, there are technical details on how we did it.
Awesome project, I just read the "About this project" section and extremely impressed! This is one of those projects I love as an adult and will be using, but I really wish I could have had this when I was a kid. Killer job!
Thanks for making this..I was trying to accurately find timestamps in a video from a transcript sometime back, was not aware that this is called "phoneme alignment" and whisperx already solves this !
One thing occurs to me about the debate between apps and web-apps. This site is really cool, but it also seems likely to be one of the endless semi-hobby sites that you check it out again after a few years and it's 404. The one thing nice about 'offline' apps is that they're always there. It'd be amazing to have a downloadable version of this. It seems it should be doable, excepting bandwidth costs if it takes off!
Unfortunately it’s almost a TB of video. The project is opensource so you could run the pipeline to grab all the videos yourself. https://github.com/bfeist/ask-an-astronaut
Yes, I was going to guess semantic embedding from playing around and seeing the search responses. Thank you for "About this project".
I'd love to extract just the foreign language from Pimsleur foreign language lessons, for faster review before trips. Your pipeline has many elements in common for what would be required.
YouTube PM if you are seeing, this is your "AI project".
Instead of just filtering user asked questions. We could probably just use an LLM to find interesting sections in video which has answer to a particular question.
I really really really love the implementation here, super fast search, amazing filtering and showcasing the video, definitely will address the curiosity of young (and old like myself) .
Congratulations on the amazing delivery. Truly great solution
I was fortunate enough to be chosen to talk to astronauts on the ISS from the ground and ask them a question, and as far as I can tell from searching this awesome resource, I think my question was unique.
I asked Ann McClain and David Saint-Jacques what experiment or module they would add to the ISS if they could pick anything. They both agreed that a rotating wheel that simulated gravity would be helpful for experiments and quality of life.
All the questions about sneeze are either about open air, or Newton's third law. Did really no one ask what happens if you sneeze while wearing a spacesuit helmet? Does the visor have a windshield wiper on the inside?
Possibly. The translation aspect makes the timing a bit more complex to work out. When I was building this out, I considered English only to be version one.
This is wonderful, thank you. A suggestion, if it's possible: Identify the astronauts speaking. (I'm not sure that can be automated reliably unless it's already in video metadata.) Also, the transcript isn't keeping up and skips parts.
I would love to, but the video metadata is entirely lacking. All I have is a filename and an uploaded date—not even an accurate date of when it was recorded.
If you select “About this project”, there are technical details on how we did it.
I'd love to extract just the foreign language from Pimsleur foreign language lessons, for faster review before trips. Your pipeline has many elements in common for what would be required.
Instead of just filtering user asked questions. We could probably just use an LLM to find interesting sections in video which has answer to a particular question.
Congratulations on the amazing delivery. Truly great solution
I asked Ann McClain and David Saint-Jacques what experiment or module they would add to the ISS if they could pick anything. They both agreed that a rotating wheel that simulated gravity would be helpful for experiments and quality of life.
But those are minor issues. Thank you so much!