Very interesting discussion here, and many great comments on p2p.

My question would be: Will media-intensive, advertising-driven
business models like Joost and Spotify (similar to Joost, but for
music) prevail in the near future or will browser-based centralized
services make for too hard competition (tv-links.co.uk, last.fm, etc)?
Bandwidth is still not very cheap when you push 100:s of TB:s/day --
and the service with the least intrusive ads wins the users in the
long run, right? So from an economics standpoint--disregarding the
fact that people have to dl specific app:s rather than simply use the
browser, etc--p2p wins in these cases, no?

From a UI point of view however, technologies like Flash (latest) and
Silverlight provide all components necessary to replicate the UI:s of
the above-mentioned p2p services (full-screen mode is the latest
addition). The only thing they can't do is save bandwidth costs
through p2p-caching. if they could, p2p would probably be used *a lot*
more than it is today.

So wouldn't a solution here be to have a generalized p2p caching layer
in the browser (with proper domain-based privacy levels, etc)? It
could be based on bittorrent or a similar technology, but would have
to be further abstracted and transparent in order to be user-friendly
enough to compete wt existing services. Maybe the WHAT-WG and/or W3C
is already thinking about this? It would certainly help the
20-something-kids-on-rails..
Or will CDN:s be able to further lower prices by pushing their
architectures in some new way? The same goes for storage (S3 is
setting a new standard here, but is still centralized).

So the second question is then: Isn't the "failure" of p2p adoption
really more a question of the browser increasingly becoming the
dominant UI paradigm vs. the older, separate apps/desktop paradigm--in
effect preventing a higher use of p2p-technology?

Regards,
eric

--
http://strategyandinteraction.com/
http://eric.wahlforss.com/