Chart of the Day

1. Key Themes

Theme: AI adoption in math research is accelerating sharply

AI acknowledgment in math preprints has hit an inflection point

• The chart tracks the monthly share of arXiv math preprints acknowledging AI use, and the line stays near zero through 2025 before turning sharply upward in 2026.
• Quote: "AI's footprint in math research is growing fast!"
• Quote (chart callout): "1 in 4 math preprints now acknowledge AI use" (the "any purpose" series reaches 25% by the final data point, around Aug/Sep-26).

Usage is moving from peripheral to substantive

• A second series tracks AI use for substantial research, not just incidental help. It climbs from roughly 0% to about 6%, so roughly 1 in 16 math preprints now credits AI with meaningful research contribution.
• Quote (chart legend): "Acknowledge AI use for substantial research" (6% at the latest data point).
• Quote (chart legend): "Acknowledge AI use for any purpose" (25%).

Growth is exponential rather than linear

• The "any purpose" line moves from low single digits in early 2026 (about 2-4% in Jan-Apr) to ~25% within a few months. The substantive-use line follows the same shape with a lag, which suggests deeper use may keep rising as capabilities improve.
• Quote (chart axis/timeframe): "Monthly share of arXiv math preprints acknowledging AI use" (Jan-25 to Jul-26+).

2. Contrarian Perspectives

The headline number overstates deep AI contribution

• The chart shows a large gap between the two series: 25% acknowledge AI use for any purpose, while only 6% acknowledge it for substantial research. Most "AI use" is likely auxiliary (editing, literature search, coding help), not original mathematical discovery. The 25% figure may reflect norm-setting around disclosure as much as capability.
• Quote: "1 in 4 math preprints now acknowledge AI use" versus the chart's "6%" for substantial research.

Measured use likely understates true use

• The data counts only papers that acknowledge AI use, so it is a lower bound on actual adoption. Math is typically seen as a field resistant to AI because of its demand for rigor, yet it shows rapid disclosed uptake. If disclosure norms lag behavior, the true penetration is higher than the chart shows.
• Quote (chart title): "Monthly share of arXiv math preprints acknowledging AI use"

3. Companies Identified

arXiv

• Description: Free, open-access preprint platform for scientific papers.
• Why mentioned: It is the data source population for the chart and a proxy for where research activity occurs.
• Quote: "A free, open-access platform where researchers share scientific papers"

Epoch AI

• Description: AI research organization that tracks AI trends and capabilities.
• Why mentioned: Cited as the source of the data behind the chart.
• Quote: "Source: Epoch AI."

Coatue Management

• Description: Investment firm publishing the "C:\Takes" newsletter.
• Why mentioned: Author and curator of the chart.
• Quote: "Chart of the Day"

4. People Identified

No individuals are named in the article.

5. Operating Insights

Treat AI-assisted research as a disclosed, emerging norm

• Teams in research-heavy fields (quant, biotech, deep tech) should expect AI-use disclosure to become standard, and should build internal policies for attribution now rather than later.
• Quote: "1 in 4 math preprints now acknowledge AI use"

Differentiate between AI as a tool and AI as a contributor

• The gap between 25% and 6% suggests competitive advantage lies in moving from incidental use to substantive integration into core R&D workflows.
• Quote: "Acknowledge AI use for substantial research"

6. Overlooked Insights

Math as a leading indicator for other rigorous disciplines

• Math is a field where correctness is verifiable and where AI errors are easily caught, making it an early proving ground. The pattern here (near-zero adoption, then a sharp inflection) may preview adoption curves in other formal domains such as physics, theoretical CS, and engineering.
• Quote: "AI's footprint in math research is growing fast!"

Timing of the inflection

• Both series flatten through 2025 and break upward in 2026, with the sharpest rise in the most recent months. This coincides with the era of newer reasoning-capable models, though the article does not state a cause.
• Quote (chart axis): "Jan-25 ... Jul-26" with the steepest slope at the right edge of the chart.