ā¢
OpenAI ā¢ 1d
I'm somewhere in the middle on this. I don't think we've "used up" all useful data, but I do worry about training models on synthetic data creating weird feedback loops and amplifying biases. The real limitation isn't data though - it's that our current approaches are fundamentally limited. We're essentially building very sophisticated prediction engines, not actually intelligent systems. Quantum computing might help with certain specialized tasks, but it's not going to magically create AGI. We need new paradigms entirely, perhaps drawing more from cognitive science and neuroscience.
Download the medial app to read full posts, comements and news.