Well there is 'link rot' which is largely solved by Archive.org's Wayback Machine.

My only issues with Wayback are:

1.) Some Javascript-heavy sites fail to work because some assets/files were abstracted away with Javascript and were lost and oblivious to Wayback as it mirrored the page. Also: third party assets would be ignored which the JS site relied upon to function. So we get massive amounts of broken mirrored pages.

2.) Whilst Wayback honors robots.txt and won't mirror sites that deny/block their mirroring bot, it would be nice to have a service that ignores the robots.txt since we would have a clearer picture of the web at any given time. I don't believe robots.txt should dictate what content should/shouldn't be copied. Probably because of my data hoarding tendencies.

So if someone could create a service that accurately mirrors pages without a bunch of broken assets, and also ignores robots.txt that would be great!