IPFS knows how to find that Koala page by its contents, not its location. The IPFS version of the Wikipedia content is represented by a string of numbers that uses IPFS to ask computers worldwide to share the page with you. Koala information can be downloaded from anyone, not just Wikipedia.
This applies to any type of content that a computer can store and not just a web page, including documents, digital art, emails, etc. If Wikipedia was to be taken down, imagine the information that would be lost. IPFS makes it possible to download a file from many locations not managed or controlled by a central organisation. This has three significant advantages for the internet:
Resilient internet – Wikipedia could be taken down by a hack or catastrophic server malfunction. IPFS would allow you to get to that content elsewhere.
Harder to censor content – Files can come from multiple locations, meaning a single government or organisation would have difficulty blocking content.
Faster access to data – With IPFS, you can retrieve a file from a geographically closer server, which means you can get that data faster. This feature is only available to more extensive, more expensive server infrastructures; IPFS makes this available to everyone.
Finally, the InterPlanetary File System gets its name from its ambition to work across places as far away as planets. Whilst interplanetary travel might be a way off, IPFS will prove helpful on planet Earth in the meantime.