In this interview, David discusses the relationship between quantum computing and artificial intelligence (AI). He believes that quantum computers will not play a central role in AI or artificial general intelligence (AGI), but rather have specialized roles in areas such as drug design or game playing. He explains that AI and AGI are fundamentally different from each other, with AI being focused on maximizing objective functions and being obedient to programmed instructions, while AGI should behave in a way that cannot be specified in advance and should provide unexpected answers. David also discusses the differences between human thinking and AI, emphasizing that humans work by understanding and conjecturing, while AI works by pruning possibilities and following a program. He believes that understanding is closely related to explanation, which involves explaining how invisible things affect visible things. He argues that current AI cannot create anything new and that any discoveries made by AI are based on pre-programmed knowledge. He also expresses skepticism about AI's ability to generate explanatory knowledge or discover new laws of physics, suggesting that these cases are likely the result of pre-programmed knowledge or data. Key themes: 1. The limitations of AI in discovering new knowledge and laws of physics. 2. The importance of moral behavior in AI and the challenges of AI making its own moral decisions. 3. The need for caution and supervision in the development and use of AI until the technology has matured. Suggested follow-up questions: 1. How can AI be programmed to make moral decisions without relying solely on pre-programmed rules? 2. What steps can be taken to ensure responsible development and use of AI in various domains?
I'm a big fan of David, but in this video he seems to be talking about AI in the old way, referring to programs and programmers, rather than how large language models work?
Oooh thank you for digging this up!
When was this recorded
Prof Deutsch is a legend. Subscribed !
In this interview, David discusses the relationship between quantum computing and artificial intelligence (AI). He believes that quantum computers will not play a central role in AI or artificial general intelligence (AGI), but rather have specialized roles in areas such as drug design or game playing. He explains that AI and AGI are fundamentally different from each other, with AI being focused on maximizing objective functions and being obedient to programmed instructions, while AGI should behave in a way that cannot be specified in advance and should provide unexpected answers. David also discusses the differences between human thinking and AI, emphasizing that humans work by understanding and conjecturing, while AI works by pruning possibilities and following a program. He believes that understanding is closely related to explanation, which involves explaining how invisible things affect visible things. He argues that current AI cannot create anything new and that any discoveries made by AI are based on pre-programmed knowledge. He also expresses skepticism about AI's ability to generate explanatory knowledge or discover new laws of physics, suggesting that these cases are likely the result of pre-programmed knowledge or data.
Key themes:
1. The limitations of AI in discovering new knowledge and laws of physics.
2. The importance of moral behavior in AI and the challenges of AI making its own moral decisions.
3. The need for caution and supervision in the development and use of AI until the technology has matured.
Suggested follow-up questions:
1. How can AI be programmed to make moral decisions without relying solely on pre-programmed rules?
2. What steps can be taken to ensure responsible development and use of AI in various domains?
🙌🏻AGI CANNOT BE A SLAVE 😊🙌🏻✨
I'm a big fan of David, but in this video he seems to be talking about AI in the old way, referring to programs and programmers,
rather than how large language models work?
It’s really the same argument though.
12:00 Ai crates new implications of existing knowledge
more computing power doesnt necessarily means more intelligence.