Hey, Skynet dev here. There a couple of major differences between IPFS and Skynet. One of the first major differences is performance. For IPFS to be high performance, it really depends on a cacheing layer. Typically, if you are not using the cache a file download can be 30+ seconds.
A second major difference is performance and scalability. IPFS really struggles to do large numbers of files, and IPFS really struggles to support large files. A single Skynet node can happily manage over 1 million files, and over 100 TB of data. Files on Skynet can get as large as 100 GB each, and TTFB will still be well under 1 second for any range request within the file.
The third major difference is how well supported dynamic content is on Skynet. Skynet has a system called ‘SkyDB’ which is like a souped-up version of IPNS. Updates to SkyDB typically take less than 200ms, and lookups on SkyDB typically take less than 100ms. You can read more about it here: Resolver Skylinks - Skynet Developer Guide - SkyDB is the main feature that enables most Skynet applications today, including our network-wide decentralized identity system MySky.
Finally, Skynet pins files to the Sia network, which means that someone using Skynet does not need to run a node all the time in order to keep the data online. You can start your node, upload some data, then turn off your node and the data will remain available. Pinning things to Skynet is also much cheaper than IPFS - it’s on the order of $5 / TB / Mo to pin things to Skynet. Further, data is stored on Sia with between 3x and 10x redundancy, depending on the type of file. This gives you much better reliability. Skynet has a much, much lower rate of data rot and disappearing files than IPFS. Which again is owed to the fact that Skynet stores data on professional hosts rather than on an ad-hoc storage network.
If you have any more questions I’m happy to help you get started.