Facebook PixelPodmine: a platform for podcast summaries
Brainstorming
Brainstorming
Idea

Podmine: a platform for podcast summaries

Podmine: a platform for podcast summaries

Image credit: Abhilash Sahoo from pexels.com

By Subash Chapagain on Oct 02, 2020

Anja M a month ago
Definitely supporting the idea.
Although I am still waging the pros and cons of human vs automated summaries. The latter is better for the reasons Juran said, and it is probably faster, but the former is better if there would be a team of people already in a couple of topics and into podcasts. This would mean that not only they would make better summaries, but also they would need less time and would know on what to focus their attention on. Additionally, if it's a job that pays, it is good for a small financial bliss of those people, as well. :)
However, for whichever option we decide, I think it's also good to make certain sections of the audio/video with a short explanation of the part and a division well-done, so that if someone is: not as interested/does not have time/can follow that particular part withouth needing much of the previous information in the video, can skip and listen to that. Additionally, it is good if you need a reminder of something particular from the podcast.
Juran K. 2 months ago
I fully support the idea, because I often listen to really long and unuseful podcasts. What I see as the problem could be the platform practicality. It would be a bit clumsy to use one platform to search and listen to full podcasts (assuming all podcasts are under some sort of protection), and the other one to search for podcast summaries. It could work out with summaries containing direct links to the original podcasts, but it would still be much easier if you have it in one place. That is why maybe it would be easier to develop an amazing podcast summarizing tool and sell it to Google Podcasts or Apple. Considering the process of summarizing itself, I think it should be AI-powered, with speech-to-text translation, keywords recognition, and listening statistics - all automatic. Statistics would contain information on whether people listen to some podcasts and podcast parts more than once (also distinguishing if it is very interesting or just badly taped). It would then collect terms and contexts used in podcasts and maybe also suggest to multiple anonymous experts a set of podcast transcripts to make a review podcast (like in science). It would also be very useful to add a function to tag certain parts of the podcast (maybe even the simplest tags like useful, boring, exciting, questionable, fake-news, etc.), like Darko mentioned. It would help to generate podcast statistics and improve the search and summarizing algorithms. The reasons why I think the mostly automated summaries would be better than manual ones are the same as Shubhankar introduced - subjectivity and long-lasting.
Subash Chapagain 2 months ago
Brilliant suggestions!
Shubhankar Kulkarni 2 months ago
Multiple people may extract multiple meanings and construct diverse useful bits of interpretation from a single podcast. A podcast can, therefore, be able to be summarized by multiple people. The summaries can then be rated/ voted and sorted accordingly.
Darko Savic 2 months ago
It sounds very useful. Many would surely need it. An additional element to incorporate is dividing the podcasts by timestamps. So that people can easily know where their part of interest begins. I often see this done in the comments section of long youtube videos. Some good person summarizes and timestamps the video.

Related ideas