[Din] Re: Fwd: New Version Notification for draft-mcfadden-consolidation-taxonomy-05.txt

[Jens Finkhäuser <[e-mail]>]

Hi all,

I feel bad for not having participated in the conversation leading up to this year's DINRG, in particular because I really like this work and the related draft on consolidation effects.

I need to reiterate, of course, that this is purely based on what was being presented. I need to read and comment on the drafts in a lot more detail.

However, just to get some conversation started:
It may be worth talking to HRPC. It's not that they have much to say on the topic of consolidation/centralization in principle. That said a number of the protocol considerations they mention in RFC9620 relate to human rights violations that are enabled by power imbalances in the digital systems humans use. Section 4.19 is explicitly about this topic ( https://www.rfc-editor.org/rfc/rfc9620.html#name-decentralization ), but e.g. the section on censorship resistance is particularly relevant due to the effective power platform providers have in more centralized systems. I suspect referencing such relationships in that RFC in the effects draft could strengthen the argument that it is not so much economic but power incentives that drive centralization.

I understand the argument that money is power and power makes money (TL;DR), but the counter argument to this is that money is the means by which power is exercised and its distribution controlled. The underlying principle of the two then cannot be the means.I spoke to Mark briefly after the session to highlight that there are also various conversations going on about "digital sovereignty", "public digital goods" and a "digital commons". They are not interchangeable, but closely related. Also, they are political terms related to similar concerns we've discussed today at the more technical level.

The differentiation between a public digital good and the digital commons is a little too subtle to go into here IMHO, but a "commons" is fundamentally a distributed/decentralized thing. The concept exists and has existed for ages, and is a tool for leveling the playing field between those with more vs. less means. One of the ongoing battles throughout history is to prevent those with more means to also exploit the commons more, because that is counter to its intent.

On the other hand, sovereignty (digital or otherwise) is about retaining power for self-determination, and not handing this to someone else. It does not necessarily imply that many people hand power to a central entity, but that's of course a possibility. So the relationship between that term and centralization of power should be understandable.There is a consideration IMHO missing from the taxonomy draft (which, again, I skimmed over only briefly just now), which one might want to call "algorithmic centralization". The specific issue here lies with the use of large language models, but also similar technology, to a) train on available (human or organically generated, etc.) data, and b) then synthesize a lot more data that is sufficiently indistinguishable from its source to effectively control the discoverability of the original source data.

There are arguments to be made that this is a subset of application/service consolidation. It's also arguable that this is less about the service or application, because what troll farms (disinformation generators) have successfully shown is that you do not need to run a service to control narratives; you need only to flood other services with disinformation.

Another argument can be made that there is a distinction between what is being generated (data consolidation) and how it is being generated (algorithmic consolidation). The use of algorithms here yields a significantly higher return on investment in terms of achieving leverage over the data space than could be gained by traditional means. Note that this, too, overlaps with HRPCs concerns about "negative impact" as mentioned in their section 4.21.

As it so happens, today the author Charles Stross happened to complain about this very thing: https://wandering.shop/@cstross/113436422346204178
"In 20/20 hindsight, I should have paid more attention to the drop in traffic to my blog (both views and comments posted) since January 2024, and noted Google's de-emphasizing blogs in favour of extruded text product, and realized: if you suppress public speech via small private media (blogs) you're implicitly boosting the reach of centrally controlled platforms.

This is censorship with American characteristics.
The fix, as they say, is in."
None of these remarks are meant to be final or replace a deeper review and/or text suggestions. It's more that I wanted to show my appreciation for the effort put into the drafts by (hopefully) sparking some discussion.

Jens

On Oct 21, 2024 at 5:30 PM, Mark McFadden <[e-mail]> wrote:
A new version of Internet-Draft draft-mcfadden-consolidation-taxonomy-05.txt
has been successfully submitted by Mark McFadden and posted to the
IETF repository.

Name:     draft-mcfadden-consolidation-taxonomy
Revision: 05
Title:    A Taxonomy of Internet Consolidation
Date:     2024-10-19
Group:    Individual Submission
Pages:    16
URL:      https://www.ietf.org/archive/id/draft-mcfadden-consolidation-taxonomy-05.txt
Status:   https://datatracker.ietf.org/doc/draft-mcfadden-consolidation-taxonomy/
HTML:     https://www.ietf.org/archive/id/draft-mcfadden-consolidation-taxonomy-05.html
HTMLized: https://datatracker.ietf.org/doc/html/draft-mcfadden-consolidation-taxonomy
Diff:     https://author-tools.ietf.org/iddiff?url2=draft-mcfadden-consolidation-taxonomy-05

Abstract:

   This document contributes to the ongoing discussion surrounding
   Internet consolidation.  At recent IETF meetings discussions about
   Internet consolidation revealed that different perspectives gave
   completely different views of what consolidation means.  While we use
   the term consolidation to refer to the process of increasing control
   over Internet infrastructure and services by a small set of
   organizations, it is clear that that control is expressed through
   economic, network traffic and protocol concerns.  As a contribution
   to the discussion surrounding consolidation, this document attempts
   to provide a taxonomy of Internet consolidation with the goal of
   adding clarity to a complex discussion.

The IETF Secretariat

_______________________________________________
Din mailing list -- [e-mail]
To unsubscribe send an email to [e-mail]