Raw LLM Responses

Speaking of AI, Ramesh's points seem to "go off script" a little bit probably because there isn't really a coherent overall message about the subject which is what we actually need for AI. I think Ramesh is more thinking about the details and has really good intuition but isn't using "big picture" thinking when he says this stuff. But his conclusion: "As new technologies create new innovation we need to create mechanisms for everyone to benefit from them" is awesome and clearly why we all consider Ramesh a progressive. There has been various ideas that have been floated about people owning their data and thus being compensated when it's used for AI like this. But for one this is obviously very capitalist. It's also not going to work because the whole reason why capitalists win isn't just "mining natural resources" like our personal info but being in a position to exploit asymmetric knowledge. You'd have no idea what the value of your personal data is why massive tech firms would. You'd always be getting low ball offers for your data. So the only correct way of dealing with this is for the intention to be that our data being applied to datasets is not accessible paid for or not. Instead any data we contribute to an AI would be voluntary and explicit in the pursuit of creating a system with intentional outcomes. Instead of a "data market" that would be a disaster (albeit a better one than the current disaster) we should be intentionally creating the data for these systems and all have input on it. THAT should be compensated. You might pay a bunch of contributors to tag photos, those photos get analyzed for systemic bias of their contributors, the sample is filtered for the specific task we want the AI to do. All of those things would be things that can be compensated. The resulting dataset is used for an AI. Inconspicuous involuntary arbitrary data collection should be banned and we don't want to continue economically incentivizing it.