Is it good that the upload bandwidth is fat?

(2024-09-16 probably wrong question)

Summary of discussions dispersed over a period of approximately one week from 2024-09-06

First thought A

• P(self): fatter upload bandwidth is self-evidently a good thing

Rebuttal B

• Q: I don’t think so.

Verbal Consultation C

• P: I got the opinion that B. What do you think?
  • R: The more perspectives we gather, the less we overlook
  • P: Well, it’s a [Lots of eyes make it easier to identify bugs.
    • No bug is too serious as long as there are enough eyeballs.

Contemplation D (by P)

• I had an unexpected answer during consultation C. I thought “I see” and didn’t confront it with my own thoughts.
• When I look back at my idea A again, I think the subject I was thinking about, the focus, was a little different.
• Homo sapiens individuals that do not upload data on their situation to the AI when the AI is distributing resources are invisible to the AI
  • It’s invisible, so it’s as if it doesn’t exist.
  • Difficult to take into account the invisible
  • If you want to be considered, it is reasonable to upload, and most reasonable people should choose to do so
    • Some may not choose to, but that’s the same genre of question as “Are you free to smoke cigarettes even if it’s hazardous to your health?” or something like that.
      • the right to execute a fool
• The more you upload, the more you gain, which creates a desire to “upload”.
  • It is not total optimization for this desire to be thwarted by “narrow bandwidth uploading means”.
  • So wider upload bandwidth is a good thing.

Reflection E (by P)

• Again, looking back at the log of rebuttal B, I thought the focus of the rebuttal was also different
• Claim E1: Increasing bandwidth will increase the amount of uploaded material and inflate the cost to the receiver.
  • My thoughts on this
    • To begin with, the receiver’s ability to receive more than ever will be enhanced by the AI, so we’ll make the bandwidth thicker to accommodate that.
• Claim E2: Some things will be “left out” in the E1 situation, and there will be an explanatory cost for what is left out.
  • My thoughts on this
    • Don’t you have to explain?
    • Once the uploads are made into open data, it becomes clear that they are enormous!
    • It becomes clear that undoing is inevitable.
    • Currently, because of the lack of open data, the requester is left with 90% of the field of view occupied by his/her own requests
      • People around me tend to have similar attributes. - filter bubble

Reflection F (by P)

• I think each of the three has a different focus to begin with.
• Is the question too abstract? I’m starting to think that it is.

My impression from reading this discussion is that it is difficult to simply judge the “good” or “bad” of bandwidth expansion, especially when the definition of “good” is context dependent. I will add my perspective on the following points

• Diversity of Purpose: When asking the question, “Is bandwidth expansion a good thing? For example, it may be “good” for the uploader in terms of convenience and improved access, but for the receiver or administrator, there is another issue of increased cost and load. Which stakeholder is central to the decision of good or bad has a significant impact.
• Technological evolutionary process: The perspective of AI expanding the capacity of receivers is very interesting, but for this to become a reality, technological development and infrastructure will need to catch up. At this stage, the increase in data due to bandwidth expansion may cause problems for current technological capacity. In other words, whether bandwidth expansion is a good thing “now” depends on the current state of the technology and how acceptable it is.
  • information asymmetry and filter bubbles: The problem of filter bubbles may be further reinforced by increased bandwidth. As open data becomes more prevalent, the bubble may widen with the increase in data volume, but at the same time, there is a risk of reinforcing the bias of the information that each individual receives. Whether this leads to overall optimization depends on how information is handled.
• The Problem of Abstraction: Finally, it is important to point out that the question is too abstract. When asking the “right or wrong” question of bandwidth expansion, the discussion would be better aligned with the specific context (e.g., individual privacy, corporate costs, optimization of information distribution for society as a whole, etc.), so that a clearer conclusion can be reached.

In short, it is my opinion that while thicker upload bandwidth may be “good” under some conditions, it is not always good from all perspectives and in all situations.

---

This page is auto-translated from /nishio/アップロード帯域が太くなることは善か？ using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.