AI Is Breaking the Moral Foundation of Modern Society

So as a preface, I would argue than in the marketplace for information (ideas, content, works, etc if you will) that value has two components, 1) utility value, and 2) scarcity value (which I would argue has two components, availability value (information horizon[0]) and saliency value (preference as a result of the information horizon[1]). Obviously the rise of copyright over the last few centuries attempts to limit commerce of ideas based somewhat loosely on the concept of (real, physical) property law (itself a glorified codified extension of the animal behaviour of territoriality, aka survival by means of securing resources); ownership is a moral value autogenous to an idea's sui generis expression on, by, or through a recorded medium. In other words the mere act of recording an idea in a "creative" fashion brings with it a "moral" ownership of that idea. Of course the "creative" and "moral" parts from which the law prescribes and proscribes these limits is debatable. The legally mandated "monopoly" of an idea (even though the marginal cost of replication approaches zero) provides this scarcity. So in the end because of this the scarcity, and the resulting "rent seeking" (corporations seeking to charging (typically over time) more than the cost to create and distribute information), are in effect value (read wealth) redistribution.

So, what is more "moral"? Using a constructivist approach, which is more socially acceptable (and hence far more likely to be codified into law)? broad access to information distilled to the point that sui generis is "removed" so that concepts (which in general can't be copyrighted) are freely available for transformative use, or extending the "moral right" of copyright, now that information distillation is so cheap, to include the ability to "copyright" an idea rather than the expression of it. Or are we seeing a "breakdown" of the codification of the law where the granularity of the spectrum is so fine it becomes so costly to enforce such the "value" of the law in reducing the transaction costs of daily social interaction in the information sphere that we revert to pre-copyright behaviours, i.e. information hoarding.

[0] https://www.researchgate.net/publication/261773523_Informati... The concept is quite simple there is a limit to the amount and quality of information that a) a human can get access to (and following what they prefer / value), and b) relatedly the amount that can be synthesised for actual use. [1] https://www.sciencedirect.com/science/article/abs/pii/S07408...

If you're doing work in The Cloud and the servers are in America, you can still move all that to physically reside in whatever jurisdiction you live in. The tech is there.

But if you build stuff on top of Anthropic, OpenAI or Google AIs, their data resides and will always reside in the US. They have very specific laws that allow them to grab data from anywhere and nobody can say shit about it.

There's no realistic way for anyone to run "ChatGPT" at home. You can get pretty close with the 200B++ param models but they're still just a pale shadow of what an acre of industrially cooled compute will do in comparison.

But if your risk modelling doesn't flash a big red light about building a business on top of someone else's hardware and software that is physically in a currently very unstable and legally unreliable country, that's fine.

I've been using my VC subsidised LLM credits to build local-first systems I can still use when either the money runs out or the political craziness crashes down.