Some of the points he raised seem right for startups to build something upon.
Publishing Platforms: Rather than rely on sites like istockphoto, photographer should start to build relationships directly with their fans and community. A open source publishing platform like BandCamp will be needed. Photoshelter is a good start but I feel it is not geared sufficiently towards helping photographers better connect with their community, such as the lack of viral distribution and analytics capabilities.
More importantly, an open source platform will allow other developers contribute to make the platform better for everyone. Photoshelter can never beat an open source version where new add-ons or new features are contributed by the community.
Ecommerce Platforms: Photographers must begin to better understand the context of where and how their images can be used. A contextually sensitive ecommerce platform can help photographers remix and rebundle their photos into different context (powerpoint backgrounds, wallpapers, scrapbook items, greeting cards), different topics (locations, celebrities, indutries), different colours et al.
The platform should allow photographer to decide how they want to sell their photos. They can sell it as a subscription, per download, or even free if the user sends in their emails. All these will allow photographers the flexibility to decided on the best way to engage their community.
Discover platforms: If photographers start to host their own sites, then there will opportunities for discovery services to help users discover new images and photographers. In the music business, we see discovery services such as Hype Machine. We need an equivlent in the photography business but using Flickr's 'interestingness' alogorithm to surface the most interesting and breath-taking image.
Related:
Startup Idea #94 (Part 1): Helping independents to earn a living