Protocols, Not Platforms: A Technological Approach to Free Speech

Notes

Some have argued for much greater policing of content online

NOTER_PAGE: (7 . 0.34214285714285714)

companies are increasingly investing in more and more sophis- ticated technology help, such as artificial intelligence, to try to spot contentious content earlier

NOTER_PAGE: (7 . 0.4114285714285714)

change Section 230 of the Communications Decency Act (CDA), which gives platforms a free hand in determining how they moderate

NOTER_PAGE: (7 . 0.475)

there should be no moderation allowed at all—at least for platforms of a certain size— such that they are deemed part of the public square.

NOTER_PAGE: (7 . 0.5164285714285715)

build protocols, not platforms

NOTER_PAGE: (7 . 0.7942857142857143)

In the past few decades, however, rather than building new protocols, the internet has grown up around controlled platforms that are privately owned.

NOTER_PAGE: (8 . 0.2357142857142857)

a single entity controlling a platform can then profit off of it. In addition, having a single entity can often mean that new features, upgrades, bug fixes, and the like can be rolled out much more quickly

NOTER_PAGE: (8 . 0.32285714285714284)

not an either/or choice here between platforms and protocols but rather a spectrum

NOTER_PAGE: (8 . 0.4714285714285714)

Historically, the internet moved more and more to a world of central- ized platforms over decentralized protocols due, in part, to the incentive structure under the old internet. Protocols were difficult to monetize.

NOTER_PAGE: (9 . 0.1742857142857143)

As it was a decentralized set of protocols, there was an involved consensus process that required agreement from a wide range of parties before any changes to the protocol could be implemented. Even smaller changes often required considerable work and even then were not always recognized universally.

NOTER_PAGE: (12 . 0.2792857142857143)

tend to be accessed through a traditional internet web browser or, increasingly, a mobile device app

NOTER_PAGE: (13 . 0.65)

this is an important point - apps lack the natural openness of the Web

it is still not clear that any platform can actually do a “good” job of moderating content at scale.

NOTER_PAGE: (14 . 0.45285714285714285)

a system of protocols could solve many of the problems associated with platforms today and that it could be done while minimizing the problems that were inherent to protocols a few decades ago.

NOTER_PAGE: (16 . 0.56)

many competing interface implementations

NOTER_PAGE: (16 . 0.7985714285714286)

Neither party may be “wrong” per se, but leaving it up to each platform to adjudicate such things is an impossible task

NOTER_PAGE: (19 . 0.12714285714285714)

increasingly complex bodies of internal “law”

NOTER_PAGE: (19 . 0.215)

A protocol-based system, however, moves much of the decision mak- ing away from the center and gives it to the ends of the network

NOTER_PAGE: (19 . 0.6028571428571429)

The key would be making sure that the “rules” are not only share- able but completely transparent and in the control of any end user

NOTER_PAGE: (20 . 0.38285714285714284)

This would not entirely prevent anyone from using the platform from speaking, but if the more popular interfaces and content moderation filters chose, entirely voluntarily, not to include them, the power and impact of their speech would be more limited.

NOTER_PAGE: (20 . 0.7514285714285714)

But we need to accept that abusive content cannot be simply "taken down" without a central authority. Of course, this is true in practice even with a central authority, but people don't like to think about it. Deplatforming vs censorship

Rather than today’s centralized system, where all voices are more or less equal (or completely banned), in a proto- col-focused world the extremist views would simply be less likely to find mainstream appeal.

NOTER_PAGE: (21 . 0.7321428571428571)

just as the filtering decisions could move to the end, so too might the data storage. While this could develop in many different ways, one fairly straightforward method is that end users would simply build their own “data stores” via apps that they control.

NOTER_PAGE: (22 . 0.17142857142857143)

much stronger incentives for the interface providers to respect the privacy wishes of the end users, as their money is likely to be driven more directly by usage, rather than by monetizing the data

NOTER_PAGE: (23 . 0.6035714285714285)

building and maintaining a protocol has long been a struggle. Most of the work was usually done by volunteers, and protocols over time were known to atro- phy without attention

NOTER_PAGE: (25 . 0.15428571428571428)

the suc- cess Google and Microsoft had had with Gmail and Outlook, respectively, show that large companies can build very successful services on top of open protocols.

NOTER_PAGE: (25 . 0.5385714285714286)

a less data-intensive ad model might thrive in the world described here

NOTER_PAGE: (26 . 0.17)

ad model based on much more limited data, with a greater focus on matching intentions or on pure brand advertising

NOTER_PAGE: (26 . 0.2792857142857143)

reverse auction type of business model, under which the end users themselves might be able to offer up their data in exchange for access or deals from certain advertisers

NOTER_PAGE: (26 . 0.4957142857142857)

build a protocol that uses a cryp- tocurrency or a token that has some value attached to it, with the value of those items growing in conjunction with usage

NOTER_PAGE: (26 . 0.6478571428571429)

appreciation in value of the tokens/currency could help fund the ongoing maintenance and operation of the protocol

NOTER_PAGE: (27 . 0.30142857142857143)

key driving factor is in getting more usage to increase the value of the tokens

NOTER_PAGE: (27 . 0.8171428571428572)

align the incentives of the network itself with a financial benefit

NOTER_PAGE: (28 . 0.2592857142857143)

Platforms have historically been quite good at focusing on the user experience aspect, especially around onboarding

NOTER_PAGE: (29 . 0.7335714285714285)

having everything controlled by a single entity also leads to some clear performance boost

NOTER_PAGE: (30 . 0.23357142857142857)

platforms themselves might consider using protocols as well

NOTER_PAGE: (30 . 0.6464285714285715)

current mode of opera- tion is only going to lead to ever more pressure to “solve” what appear to be unsolvable problems.

NOTER_PAGE: (31 . 0.13071428571428573)

Well, not really - they don't need to solve these problems to keep making money, at least as long as the law doesn't care enough to make it consequential

continuing with what they are doing is going to be increas- ingly costly. Already Facebook recently promised to hire another 10,000 moderators

NOTER_PAGE: (31 . 0.2592857142857143)

and they can easily afford it, so what

competing against other large internet platforms in areas where they have much less ability to compete

NOTER_PAGE: (31 . 0.43214285714285716)

Oligopolists are unlikely to undercut each other like this

if the token/cryptocurrency approach is shown to work as a method for supporting a successful protocol, it may even be more valuable to build these services as protocols

NOTER_PAGE: (31 . 0.6257142857142857)

allowing abusive individuals—whether mere trolls or horrifying neo-Nazis—to have any ability to speak their minds is going to be a problem.

NOTER_PAGE: (31 . 0.7964285714285714)

we already have those people infesting the various social networks, and nothing so far has been successful in getting rid of them.

NOTER_PAGE: (32 . 0.19357142857142856)

while they would be able to be vile in their own dark corners, their ability to infect the rest of the internet and (importantly) to seek out and recruit others would be severely limited.

NOTER_PAGE: (32 . 0.2785714285714286)

allowing them their own little space in which to be crazy might better protect the wider internet

NOTER_PAGE: (32 . 0.4707142857142857)

There is little reason to think that in a protocol-focused world, this problem would be all that different than what currently exists.

NOTER_PAGE: (32 . 0.8407142857142857)