Viral Thread on the Accelerating Leap in LLM Capabilities
morqon · x · 2026-07-16
This retweet summarizes the pace of LLM progress as a steep curve:
- In 2023, they struggled with 4th-grade math
- In 2024, they could handle high school math
- In 2025, they allegedly won a gold medal at the IMO
- It then offers a more radical take: GPT-5.6 can already solve frontier math/statistics problems, making an IMO perfect score "no longer news"
The quoted post discusses a specific statistical result: AI helped solve a theoretical problem regarding FDR control under correlated data in multiple hypothesis testing, extending the Benjamini-Hochberg method beyond independent assumptions.
Related event: Debate Intensifies Over the Rapid Leap in LLM Mathematical Capabilities(3 posts)→
More from AGI Musings
- Gary Marcus says LLMs still cannot really do math on their own — GaryMarcus · 2026-07-22
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22
- Better AI math could save researchers time by killing false conjectures earlier — prateekj · 2026-07-22
- AI’s economic forecasts are split by nearly a quadrillion dollars by 2035 — bittingthembits · 2026-07-22
- Open source is becoming tech’s soft power, says Kevin Xu — kevinsxu · 2026-07-22