When can we have the ability to Discover/Prove New theorems ?
This is response from GPT 3.5.
ChatGPT 4.0, or any AI model, does not possess the innate ability to discover or prove new theorems autonomously. It lacks the creative and innovative thinking required for ground-breaking mathematical research. While AI can assist in theorem proving by offering mathematical insights, generating proofs based on existing knowledge, or automating certain aspects of the process, it relies on pre-existing mathematical knowledge and cannot produce genuinely novel results. Theorems, particularly ground-breaking ones, emerge from the brilliant minds of mathematicians who apply deep understanding, intuition, and inventive thinking. AI, including ChatGPT 4.0, serves as a valuable tool in mathematical research but doesn’t replace human creativity and ingenuity.
You might find this category on the Lean forum of interest.
https://leanprover.zulipchat.com/#narrow/stream/219941-Machine-Learning-for-Theorem-Proving
Also check out
