Source URL: https://simonwillison.net/2025/Sep/29/scott-aaronson/
Source: Simon Willison’s Weblog
Title: Quoting Scott Aaronson
Feedly Summary: Given a week or two to try out ideas and search the literature, I’m pretty sure that Freek and I could’ve solved this problem ourselves. Instead, though, I simply asked GPT5-Thinking. After five minutes, it gave me something confident, plausible-looking, and (I could tell) wrong. But rather than laughing at the silly AI like a skeptic might do, I told GPT5 how I knew it was wrong. It thought some more, apologized, and tried again, and gave me something better. So it went for a few iterations, much like interacting with a grad student or colleague. […]
Now, in September 2025, I’m here to tell you that AI has finally come for what my experience tells me is the most quintessentially human of all human intellectual activities: namely, proving oracle separations between quantum complexity classes. Right now, it almost certainly can’t write the whole research paper (at least if you want it to be correct and good), but it can help you get unstuck if you otherwise know what you’re doing, which you might call a sweet spot.
— Scott Aaronson, UT Austin Quantum Information Center
Tags: gpt-5, quantum-computing, generative-ai, llm-reasoning, ai, llms
AI Summary and Description: Yes
Summary: The text illustrates a practical experience of utilizing AI, specifically GPT-5, to assist in complex intellectual tasks related to quantum computing and problem-solving. It highlights the collaborative potential of AI in academic settings, showcasing its capability to refine ideas and enhance the creative process despite inherent limitations.
Detailed Description: The provided content reflects an interaction between a researcher and an advanced AI model (GPT-5) in the context of solving a complex problem in quantum computing. Key insights include:
– **AI’s Role in Intellectual Tasks**: The text discusses how AI can aid in tasks traditionally considered uniquely human, such as proving complex theoretical concepts in quantum computing.
– **Iterative Problem Solving**: The interaction with GPT-5 was iterative, illustrating how AI can improve its responses based on user feedback, akin to collaborating with a graduate student or a colleague.
– **Limitations and Capabilities of AI**: While the AI was unable to produce an entire research paper, it effectively guided the user past specific hurdles, indicating a role for AI in enhancing existing knowledge rather than replacing human intellect completely.
– **Future of AI in Research**: The narrative suggests that AI’s integration into academic research could become more widespread, changing how complex intellectual problems are approached.
This interaction underscores the emerging capability of AI in academia and the lifting of cognitive burdens associated with complex problem-solving, which could have significant implications for how security and compliance professionals in AI-driven sectors consider the integration of AI tools in their workflows. The potential of AI in supporting human intellect could also raise concerns about the accuracy and reliability of AI-generated content, necessitating robust oversight and a deep understanding of the tools in use.