As an engineer working at a startup that is actively challenging the entertainment status quo through a slightly non-traditional vector, this RFS strikes me as amazingly naive and lacks any realistic suggestion for what kind of challenges might be most effective.
Hollywood isn't dying, nor is it anywhere close. Shedding some excess, maybe, but not dying. It's an overweight person rationalizing away his weight problem instead of committing himself fully to do whatever it takes to become lean and efficient. Apologies if that offends; I don't mean any malice. I simply mean to point out that even without Hollywood, there are any number of entertainment pieces constantly being produced and distributed. Indeed, Hollywood is, in my opinion, little more than a destination and mini-society that brings together and publishes works, as well as an ecosystem of funding. Try finding an animated movie that truly, terribly falls flat like some films. It'll be harder than it would be for a fully live-action film, because they know that without star power or explosions as a fallback, that much more effort needs to be spent on narrative to produce something that, if nothing else, will not tank incredibly. Is it any wonder that the animated versions of large or popular franchises typically have terrible movie productions†?
(† in this case I am also including live action movies with a very major animated component, i.e. a mix of live-action and animated characters)
The people who build the technological platforms hosting today's entertainment are, by and large, not dumb people. It takes some real skill to scale a site that serves up any number of combinations of video profiles and different target platforms. This, especially, is somewhere the cloud has trouble with cost-effectiveness. Video loads are large, and can be CPU, RAM, and I/O bound (if you're trying to make the most of your resources), and iteration and improvement on the process can take hours or days, not simply minutes, for almost-imperceptible but possibly still important improvements. There are reasons it works for Netflix and reasons that it's more expensive for smaller companies to scale. There's also the fact that DRM is often a contractual obligation, and the start-up costs can run well into the several hundreds of thousands of dollars range, not including high upkeep contributing to a gargantuan TCO by any startup's standards. Then there's the revenue sharing, of course.
The problem is not technology, or throwing more startups at the problem. The innovation must come from hacking the business side, or altering the deal with how content providers bring their content to the masses. Building more technology without doing this will waste people's lives and embitter them, because entertainment is being held back not by the available channels to get it to you, but because taking risks and failing fast nets you a blacklist from the people licensing you the content. Put another way, the content providers are choking off new media, not the actual proponents of new media. The pipeline of people making a living off of this stuff leads into hollywood and anything that can be done to divert them away from that walled garden and into a content bazaar will do more to kill Hollywood than actually targeting Hollywood ever will.
The solution, in my opinion, is to empower the entertainers and the content producers to reach their audience more directly, more broadly, and with more engagement than a TV or, let's say, a comic book will ever be able to give you. As a medium for entertainment production, YouTube falls flat. It's not the product online entertainment wants to be. It's not a product designed for simulcasts, and it's not a product designed for building communities, or for social watching, or for alternative media formats, or for worldwide publishing with subtitles of customizable language or styles, with pre-roll, mid-roll, or post-roll ads, or for subscription-based monetization. And even if some of these exist separately, they're not all in one place.
Also, large video ad tech is fucking terrible and poorly specified, except for Hulu, which has full access to its own tech, which it doesn't license out, as it runs its own in-house ad network. Ad companies tend to have shitty engineers or get bought by Google. But Admob's video ads monetize poorly compared to the other (technically worse) options, surprisingly.