While imho the purpose of ipfs differs from http, I think that this project will have severe problems in the near future. To name a few
- They use SHA2-256 as the default hash algo. SHA1 is broken already, so why do they use the next one on the list of things to be broken? How will they transition to a different algo while preventing breaking everything in existence?
- For now, they only can share static files. most of the web is dynamic though, and you cannot hash dynamic content. It's against their own principles, so they will not be able to provide that at all.
- Where is their formal documentation for how it should work? Any standardization draft? The world will have a hard time accepting tools without standardization...
- Their marketing sux. I read the website and watched the video, and all I heard was: cool, I can connect and it transmits data. That's something worth an hour of work and does not explain the advantages of ipfs at all. Why should I use it at all? What can it do which I cannot do with the current web already? Where are they headed to? Why Go (they use it as a marketing argument(!!)) and why does that even matter?
I think, if you really want to use it, then you might use it in addition to http, for example as a cdn.