Terence Tao · Oct 18, 2026
Francesco Maggi: AI’s mathematical proofs could usher in a golden age—or weaken human science
Markdown · JSON · Original source
In a guest post on Terence Tao’s blog, Francesco Maggi argues that AI systems are already producing mathematical proofs worth serious scrutiny, but their results do not automatically amount to human understanding or lasting scientific progress. Some proofs written in Lean—a system that lets computers check formal arguments—are so long that mathematicians may struggle to interpret them. Maggi sees a major opportunity in using AI to help develop mathematical frameworks for studying natural and social phenomena, while warning that proof-solving alone is not the whole of mathematical progress. The outcome, he argues, depends on choices about how these tools enter research. He calls for clear credit and review practices, protection for unpublished research, training that helps students retain independent abilities, and broader access to powerful systems. His proposed response is coordinated, publicly accountable scientific infrastructure—not necessarily independence from companies, but enough shared capability for researchers to test claims and pursue their own questions. How AI affects creativity and learning remains uncertain; Maggi presents a “golden age” as achievable, not inevitable.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
Terence Tao · Article
Source details
Original source date: Oct 18, 2026 (2026-10-18T14:59:49+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-10-19T04:02:06.636Z. First observed: 2026-10-18T15:01:25.820Z. Retrieved: 2026-10-19T04:01:24.200Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The post is explicitly identified as a guest post by Francesco Maggi; Terence Tao’s note says it was converted from a different file format using AI. The post offers arguments and proposals, not evidence that its predicted benefits or risks will occur. It says the effects of intensive LLM use on creativity are largely unknown. The post describes certain generated proofs as notable examples but does not establish their correctness or broader real-world reliability in the supplied evidence.
Open question
What standards can show that AI-assisted mathematical work reflects a researcher’s understanding and contribution, rather than only a machine-checked result?
Source-stated
Guest author Francesco Maggi argues that AI-generated mathematical proofs merit serious independent examination, while their quality and limits are not yet fully understood.
While the quality and limits of these tools remain incompletely understood, their potential for accelerating discovery
Source-stated
Maggi says some AI-produced Lean proofs are extremely long and difficult to translate into natural language, making human interpretation necessary before they can contribute to understanding.
LLMs can give us correct Lean proofs that we cannot directly read, and messy natural language translations
Source-stated
Maggi argues that mathematical progress includes creating new mathematical languages and frameworks, not just proving listed conjectures.
The act of inventing the very mathematical language through which reality (including mathematical structures themselves) becomes quantitatively and rationally intelligible.
Source-stated
Maggi warns that unclear attribution, responsibility, and confidentiality could discourage researchers from sharing ideas or using AI tools.
Which mathematical ideas contributed to a specific proof? Who is responsible for asserting the proof’s correctness
Source-stated
Maggi says researchers lack satisfactory standards for assessing individuals’ mathematical contributions in the age of LLMs, and that effects on creativity are largely unknown.
we have yet to develop satisfactory standards for assessing individual intellectual contributions to mathematical research in the age of LLMs.
Source-stated
Maggi argues that unequal access to frontier systems could concentrate influence over research questions, tools, and the release of discoveries.
Allowing private interests to determine which questions receive the best tools, who can use them, how capabilities are evaluated, and when discoveries are released would concentrate scientific influence to an unprecedented degree.
Source-stated
Maggi proposes coordinated scientific infrastructure and concrete steps such as access, reproducibility standards, education research, and public investment to make a human-centered outcome more likely.
Access to frontier LLMs for researchers, reproducibility and disclosure standards for LLM-generated scholarship
Source and contributor context
Terence Tao · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media was declared by the source, but no asset passed the supported availability checks. Open the original source to view it.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Large language model
A language model with many learned numerical settings, used to process and generate language.
Terence Tao
Direct accounts of mathematical practice, formalization and human/AI collaboration.
Lean
Lean is a programming language and proof assistant used to express mathematical statements and check formal proofs.
Editorial provenance
Version 1. Publication ID: 7469dfa9-3f21-4569-871b-a4e266000b88.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-19T04:02:06.636Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 607e00a6f17c406321eaf4c5ae7ab2abcdc16ce8aaa68ee90013994c9c3fe8f8. Source text hash: 300aa92ec13dc85a03e30d5de4e70a727f2205d9c4b68dbb0fbff909b2c4dd4d.
Revision and source-check history
Version 1 · 2026-10-19T04:02:06.636Z
Francesco Maggi: AI’s mathematical proofs could usher in a golden age—or weaken human science
In a guest post on Terence Tao’s blog, Francesco Maggi argues that AI systems are already producing mathematical proofs worth serious scrutiny, but their results do not automatically amount to human understanding or lasting scientific progress. Some proofs written in Lean—a system that lets computers check formal arguments—are so long that mathematicians may struggle to interpret them. Maggi sees a major opportunity in using AI to help develop mathematical frameworks for studying natural and social phenomena, while warning that proof-solving alone is not the whole of mathematical progress. The outcome, he argues, depends on choices about how these tools enter research. He calls for clear credit and review practices, protection for unpublished research, training that helps students retain independent abilities, and broader access to powerful systems. His proposed response is coordinated, publicly accountable scientific infrastructure—not necessarily independence from companies, but enough shared capability for researchers to test claims and pursue their own questions. How AI affects creativity and learning remains uncertain; Maggi presents a “golden age” as achievable, not inevitable.
Original source
Terence Tao · Article
Source details
Original source date: Oct 18, 2026 (2026-10-18T14:59:49+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-10-19T04:02:06.636Z. First observed: 2026-10-18T15:01:25.820Z. Retrieved: 2026-10-19T04:01:24.200Z.
Limits
The post is explicitly identified as a guest post by Francesco Maggi; Terence Tao’s note says it was converted from a different file format using AI. The post offers arguments and proposals, not evidence that its predicted benefits or risks will occur. It says the effects of intensive LLM use on creativity are largely unknown. The post describes certain generated proofs as notable examples but does not establish their correctness or broader real-world reliability in the supplied evidence.
Open question
What standards can show that AI-assisted mathematical work reflects a researcher’s understanding and contribution, rather than only a machine-checked result?
Source-stated
Guest author Francesco Maggi argues that AI-generated mathematical proofs merit serious independent examination, while their quality and limits are not yet fully understood.
While the quality and limits of these tools remain incompletely understood, their potential for accelerating discovery
Source-stated
Maggi says some AI-produced Lean proofs are extremely long and difficult to translate into natural language, making human interpretation necessary before they can contribute to understanding.
LLMs can give us correct Lean proofs that we cannot directly read, and messy natural language translations
Source-stated
Maggi argues that mathematical progress includes creating new mathematical languages and frameworks, not just proving listed conjectures.
The act of inventing the very mathematical language through which reality (including mathematical structures themselves) becomes quantitatively and rationally intelligible.
Source-stated
Maggi warns that unclear attribution, responsibility, and confidentiality could discourage researchers from sharing ideas or using AI tools.
Which mathematical ideas contributed to a specific proof? Who is responsible for asserting the proof’s correctness
Source-stated
Maggi says researchers lack satisfactory standards for assessing individuals’ mathematical contributions in the age of LLMs, and that effects on creativity are largely unknown.
we have yet to develop satisfactory standards for assessing individual intellectual contributions to mathematical research in the age of LLMs.
Source-stated
Maggi argues that unequal access to frontier systems could concentrate influence over research questions, tools, and the release of discoveries.
Allowing private interests to determine which questions receive the best tools, who can use them, how capabilities are evaluated, and when discoveries are released would concentrate scientific influence to an unprecedented degree.
Source-stated
Maggi proposes coordinated scientific infrastructure and concrete steps such as access, reproducibility standards, education research, and public investment to make a human-centered outcome more likely.
Access to frontier LLMs for researchers, reproducibility and disclosure standards for LLM-generated scholarship
Source and contributor context
Terence Tao · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media was declared by the source, but no asset passed the supported availability checks. Open the original source to view it.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Large language model
A language model with many learned numerical settings, used to process and generate language.
Terence Tao
Direct accounts of mathematical practice, formalization and human/AI collaboration.
Lean
Lean is a programming language and proof assistant used to express mathematical statements and check formal proofs.
Editorial provenance
Version 1. Publication ID: 7469dfa9-3f21-4569-871b-a4e266000b88.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-19T04:02:06.636Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 607e00a6f17c406321eaf4c5ae7ab2abcdc16ce8aaa68ee90013994c9c3fe8f8. Source text hash: 300aa92ec13dc85a03e30d5de4e70a727f2205d9c4b68dbb0fbff909b2c4dd4d.
passed · 2026-10-19T04:02:06.636Z · anthropic/claude-opus-5.5
Publication: 7469dfa9-3f21-4569-871b-a4e266000b88
attribution — pass: The draft correctly identifies the piece as a guest post by Francesco Maggi on Terence Tao's blog, attributes the arguments to Maggi, and notes Tao's statement that the post was converted using AI.
[This is a guest post by Francesco Maggi. This blog post was initially written in a different file format and converted using AI. — T.]
summary — pass: The title, summary and question match the post's central argument about golden-age and dark-age scenarios. The draft explains Lean in plain language and covers the post's main points: human understanding, mathematization, credit and peer review, training, unequal access, and public infrastructure. The question follows from the post's concern about how to assess individual contributions.
The golden age scenario is achievable.
date — pass: The originalPublished value, 2026-10-18T14:59:49Z, comes from the publisher page. It matches the feed entry and the date in the URL, and there is no sign of an earlier first publication.
support — pass: Each claim is an attributed argument with an exact supporting excerpt. The warning that researchers may become reluctant to share ideas or use AI tools is supported by the post's discussion of guarded sharing and reduced LLM use. The limitations do not overstate the evidence: they treat the post as argument rather than proof.
So, another side of obscurity and uncertainty is that people may end up using LLMs much less than they would like to do.
media — pass: The selected media is the verified direct link to the original article, which is the most direct and relevant way to access the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.