Such megadumps are only really interesting for people who train LLMs anyways. No way you can ever consume that for personal development. Individual sources are much more prevalent and useful to everyday life.
Get the data, make embeddings, categorize and tag the data, then do search and have LLM summarize findings.