This is a weekly update on my progress to document the history, perspectives, and narratives of the metaverse. For more on the contents and structure of the book, please see “Structuring my field notes“.
Current book word count: 69,698
Talking
I was on vacation last week, so not much writing happened. However, I was able to meet with Thomas Riedel for a relaxed breakfast. We talked about his podcast Spatial Realities (probably the best Metaverse podcast in German), the current state of spatial computing, and why we do what we do. Thank you, Thomas, we need to do this more often😊
I am also working on the interview with Julian Dibbell, which is shaping up to be a good one. The central theme is the intersection between the legal realities of virtual worlds and their theoretical potential to be something more, or rather, something else. This already came up during the interview with Raph Koster, but this time we started with the legal practicalities and then ventured into social governance.
Other
I have been thinking.. up until now I have used generative AI on this blog. More specifically, I used images created by Midjourney to create the article features. I decided to get rid of them.
I have been vocal that while I believe that LLMs and generative AI in general have their uses, I don’t like the way many companies use copyrighted material to train their models.
I don’t buy the fair use argument, or rather I personally think that this should not constitute fair use. And if it is “impossible to be successful without violating other platforms terms of service and exploiting copyrighted material“, then, well, you should not be in business.
As a result I have removed all generative AI images from this blog and cancelled all my subscriptions with the respective services. This has no effect on the book itself, since I don’t use generative AI to write it. (I tried, but the results were ultimately not good)
For more information, see:
- The copyright conundrum of AI art – The Verge
- Generative AI Has an Intellectual Property Problem (hbr.org)
- The scary truth about AI copyright is nobody knows what will happen next – The Verge
In time, I might look for a commercial model trained on properly licensed material only.