Gilbert responded to the reactions, stating that the older generation of researchers struggle to understand that the ground has shifted beneath them, with much of the intellectual energy and funding today coming from people not motivated by science. He argued that deep learning was motivated more by the metaphor of neurons in the human brain than a clear understanding of what intelligence amounts to. Gilbert concluded by saying the overall discussion should be an invitation to think more deeply about what we want intelligence to be.
Key takeaways:
- Machine ethicist Thomas Krendl Gilbert and OpenAI co-founder Ilya Sutskever have both referred to AI as a form of 'alchemy', suggesting that the process of building AI is more mysterious and less understood than traditional science.
- Meta chief AI scientist Yann LeCun disagreed with this view, arguing that AI is a form of engineering science, not alchemy. He believes that those who dismiss AI as alchemy are misunderstanding the empirical science behind it.
- Gilbert argued that the current generation of AI researchers are less motivated by theory and more by the belief that they are creating a new era of consciousness with superintelligent machines. He suggested that this shift in motivation has led to a lack of understanding about the foundations of AI.
- Gilbert concluded that the discussion about AI should prompt deeper thinking about what we want intelligence to be, and that these are human questions, not strictly scientific ones.