December 13, 2019•673 words
Privacy activists have an alternative for every Google product except YouTube.
YouTube defines the long-form, user-generated video market. Which is troubling for privacy-activists & general folk alike. One company exerts near-perfect control over the juciest portion of the world's media diet.
Under pressure from the legacy media, company leadership favors the iron fist to the velvet glove. With a new terms of service, YouTube re-iterates it's right to not just demonetize creators, but remove them wholesale from the platform.
This is an implicit ban on politically-incorrectness. Even worse is the possibility of Google letting foreign governments censor content on American soil. Meeting international demands is expected of international companies, but countries without a Bill of Rights must not be allowed to infringe on ours.
Long-form, user-generated video is too important to leave inside a walled garden. We need an escape tunnel.
Finding Cracks In The Wall
What maintains YouTube's dominance?
I see YouTube as a nexus of 3 features:
- Hosting: content must live somewhere
- Monetization: incentivize creators to consistently produce high-quality content
- Search: connect users to requested content
- Recommendation: keep users engaged with relevant, personalized suggestions
Which feature is the wall? It's not hosting; hosting is a commodity. It's not monetization; smart creators know to diversify their revenue across sponsorships, Patreon, branded products, etc. Given the multitude of ways for creators to draw income, it's doubtful their loyalty to YouTube extends past the level of convenience & inertia. So what is the wall?
Discovery is the wall. YouTube is the only meaningful video search engine & maintains this status by never recommending videos outside itself. Viewers don't know where else to find video & creators don't know where else to find viewers. Even if a viewer deigned to use, say, DuckDuckGo video search:
the search results would map mostly to YouTube
the user still depends on YouTube for recommendations 'down the rabbit hole'
Discovery can bless content with international acclaim or doom it obscurity. Thus discovery leads, viewership & content follow.
Breaking The Wall
What if discovering video was de-coupled from hosting it?
Traditional web search is platform-independent; Google finds the content you seek & directs you to the source. A video-search/discovery service predicated on the open web could route around corporate censorship & empower creators to engage fans directly, on the creator's own site .
But there's a key difference to between web search & video discovery. Web search is like visiting the DMV; the goal is to spend as little time as possible. Video discovery is like visiting a casino; the more time spent, the better.
So as enumerated in the DuckDuckGo example above, rehashing web search as video search isn't satisfactory. A balance must be struck between displaying suggested content from around the web, while not eclisping the site which the viewer is currently on. Some sort of toggle-able overlay of suggestions, perhaps.
Divorcing discovery from hosting doesn't just grant freedom to creators - it gives options to viewers, too. If the concept works, we can assume competiting discovery services will emerge.
Users can vote with their feet as to which service 'discovers' best. As opposed to the natural monopoly over video-recommendation currently enjoyed by YouTube.
A Better Garden, Not Just A New One
We must be careful not to take one step forward & two steps back. If we are to create a new paradigm, it must be an order of magnitude better.
My personal desires are:
- privacy-respecting: user-supported rather than ad-supported
- laissez-faire regulation: any content which doesn't violate US law is indexed*
- users may block content they dislike & subscribe to common block-lists
- transparency: recommendation algorithms steer the course of public discourse. Open-source algorithms serve the greater good.
(*Except porn, which tends to swallow websites & should be treated seperately)
There's a million more details to enumerate, but it's better to be published than perfect. If you enjoy the idea, please drop me a DM on Keybase or Twitter.