Hi, this is Wayne again with a topic “What Can AI Companies Learn from Web3? | TechCrunch”.
There are two wildly underappreciated fragilities of what’s going on with AI right now in the mainstream. The first is: how were these incredible large language models built? They were built by training on most of the internet, I’m sure they scraped the TechCrunch website. They didn’t ask permission for that, and they didn’t give us any compensation. I didn’t get any open AI shares, I’m not sure if you did for all your great writing, but we all helped train these models and we got no compensation for it. That is patently unfair and possibly illegal, probably possibly a copyright violation and the courts are going to sort this out well.
This is a great example of something that web 3 has thought a lot about. How do we create systems where many many people or creators can contribute content to them in exchange for some piece of ownership of them? So we can imagine a decentralized version of open AI where all of the people whose websites were scraped got some tokens, maybe in prop poron for the amount of content that they contributed. So you can imagine getting paid in tokens and then in order to use these networks of products, you have to buy tokens and that could create a a sort of Commons based ownership model for for AI, where everyone’s getting compensated. .