In September last year, Peter Mandelson was fighting to keep his job as British Ambassador to the US after the first raft of revelations about the extent of his friendship with Jeffrey Epstein.
Within hours of the details emerging, an anonymous Wikipedia editor had made changes to Mandelson’s page that distanced him from Epstein and cast him in a sympathetic light. That editor has since been blocked for making undisclosed paid changes.
New details about the relationship between the two – including that Mandelson recommended a villa where Epstein could host his “guests” – have sparked a national scandal in recent weeks and led to pressure on Keir Starmer to step down as prime minister.
But over the course of two days in September, while Mandelson was still in his government job, the mysterious account made a series of edits that either reflected more favourably on him or pushed details of the Epstein scandal under unrelated information.
And when Mandelson was eventually sacked on 11 September, it moved within hours to remove the reason given by the Foreign Office for his dismissal: that Mandelson had told Epstein his 2008 conviction for sex offences was wrong and encouraged him to clear his name.
It’s a good thing that lots of people have full backups of wikipedia.
I saved a copy for myself at the start of 2025. It took about 23GB of space if I’m remembering right. Maybe I’ll burn a blu-ray copy for long term storage
Is that the compressed version? Kiwix’s latest copy is roughly 100GB including images.
That’s right, I forgot about that. It was 23.1 GB compressed for the “enwiki - pages-articles-multistream” version that includes pictures.
Uncompressed and set up with XOWA viewer it’s about 70 GB
I assume it was a subset of languages. Only EN is much smaller than full with all languages.
You should print it out:
Wikipedia isn’t important because of its data. Rather because of the fact it is continuously updated, extended, and fixed at a gigantic scale.
If Wikipedia ever dies, its information will lose relevance by the day. After a decade or two without a similar-scale replacement, will anyone even care?
No, the data itself is inherently valuable even when it’s a little bit dated. We don’t need daily updates to learn about historical events, methods of irrigation, 20th century election results, mineral composition of transistors and diodes, and millions of other well-documented topics. It’s an incredible resource of collected knowledge with immense inherent value.
Imagine having a collection of Wikipedia backups and disclosing that on a first date.
What are we reading today, babe? 2020q1, the Covid hoaxes? Yeah, that’s the shit.