- Elon Musk, in a conversation with Mark Penn, stated that AI training has used up the cumulative sum of human knowledge, necessitating a shift to synthetic data.
- Musk’s views align with those of former OpenAI chief scientist Ilya Sutskever, who predicted a move away from current model development due to ‘peak data’.
- Tech giants like Microsoft and Meta are already utilizing synthetic data, which offers cost benefits but also poses risks of bias and model collapse.
Elon Musk highlights the exhaustion of AI training data and the future role of synthetic data
