Thinking about it teeth What does that mean? What does it mean for humans to think? Are we being overtaken? artificial Do you think so? A lot of people do, but as far as I understand, Teppo Ferrin doesn’t think so. In this episodeEconoTalk host Russ Roberts welcomes Felin back to discuss these questions, based on a working paper he co-authored with Matthias Hollweg. Felin argues that human capacity is ignore Existing data and evidence are not only our weakness, but also one of our superpowers. do When you think???
After listening to the podcast, we’d love to hear your thoughts. Share your answers to the questions below in the comments, or use them to start an offline conversation. Either way, we’re rooting for you.
1- What’s wrong with thinking of the brain as a computer? This widespread metaphor has been around since the 50s. According to Ferrin, while it’s true that neural networks form connections, the brain is not yet fully a computer and there’s something else going on. According to Ferrin, how should we understand the brain instead?
2- Large language models can capture so much content that it would take humans hundreds of years to pre-learn the same number of words, says Ferrin. He explains that LLMs employ stochastic processes that are very good at “predicting what’s next.” If LLMs contain so much content, why can’t they think like humans? What is the importance of looking back in the human thought process?
3- Ferrin argues that using AI to make rational decisions is foolish, as is the idea that we can remove bias from AI. Why is bias a feature rather than a bug of human cognition? What is the relationship between beliefs and theories and thinking, and how do we distinguish human thinking from AI?
4- Felin gives the fascinating example of the Wright brothers and argues that venture capitalists did not fund them. According to Felin, why is AI Hindered Why did the Wright brothers make such progress? Conversely, why were they successful?
5- Roberts quotes Yann LeCun: “Prediction is the essence of intelligence.” To what extent does prediction have to do with intelligence, according to Ferrin? Ferrin says: “…that’s something computers can’t do. Computers take existing data for granted, whereas humans discover and create new things through experimentation. New DataTo what extent is there a meaningful difference between digital and biological intelligence, and how would you explain it?
