Open sourcing the training data for ELM is basically the only way to _ever_ get limited trust into AI answers. Given that neural networks are still conceptually the same thing as they were 30 years ago (which is when I happened to learn about them) my hopes are low. These things work by doing statistics nobody can understand and they will only happen to create the right abstraction, usually there will be none at all. Still the _oposite_ of intelligence and moving in the wrong direction.
AI has gotten so much better these last few months. It's so good to the point where I almost feel bad for asking ChatGPT to do so much work haha.
Open sourcing the training data for ELM is basically the only way to _ever_ get limited trust into AI answers. Given that neural networks are still conceptually the same thing as they were 30 years ago (which is when I happened to learn about them) my hopes are low. These things work by doing statistics nobody can understand and they will only happen to create the right abstraction, usually there will be none at all.
Still the _oposite_ of intelligence and moving in the wrong direction.