Using the Internet Archives open API (based on the finding of Alexa an 1996 search engine) I created a webpage where one can search for websites in an older version. Older colors, older flow, older interface and sometimes completely different purposes (like in the case of facebook.com which had a different owner in ’99 before the social network Facebook was founded in ’04).
The begging default year to travel to is 1999 but if requested website didn’t exist yet it will automatically send the user to the nearest past available.

Its important to notice NOT all existing websites and all their versions are archived because of memory constrains. One might not find a given website, or, like in my case, just a one older record of it. Give it a try on this link. Also, the code is available on my github.

Next steps:

Finding a way to make the process faster so that it possible to scroll through time and see the change happening live.
Another direction is to pull only the styles sheet (css) of a site (or of a certain time?) and apply that on whatever website chosen. You can apply old on new, new on old, or just one on another.  To make it more practical I thought of having a slider that allows to tune the amount of the source styles to apply (as some elements would brake without their css).

This work is also connected to the Time machine research Im conducting these days.