I’m on a slightly modified version of 0.52.1 which is getting a bit dated but it works well for me even with not officially supported source, like svelte.
In case this thread helps someone else, some errors with —show-repo-map can be solved by setting environment variable PYTHONIOENCODING=utf-8
For me, this has been perplexity.ai. Give it a query, it expands that into multiple queries (possibly chained depending on results) and synthesizes the results into an answer with citations.
I wondered how this could reliably distinguish between a scene cut and a cut to commercial without content hashes and/or program schedules being shared through the network, then realized that 20 years ago was already 2003 and of course home internet was common by then.
Apparently one offline technique was checking for black frames inserted by local stations.
Update: I got an automated email from Goodreads to download this export. It contains a bunch of json files containing user/usage-related data for my account like request logs, newsfeed updates, kindle logins, site settings, etdc.
I assume this data export includes much more than just the book data (which I just discovered you can get immediately as CVS via https://www.goodreads.com/review/import) which might be part of a compliance requirement.
They will send a confirmation email. After you click the link to confirm, it says "We will provide your information to you as soon as we can. Usually, this should take no more than a month."
EDIT: you can export your book database as CSV (shelves, ratings, reviews, etc) here https://www.goodreads.com/review/import (there's a link towards the top to export...disregard that the URL says "import".
Reddit is not bad for discovering content if you understand the inherent biases, and I personally have good results with blogrolls on blogs I already enjoy such as https://slatestarcodex.com/.