OpenAI reports superexponential van der Waerden lower bounds

There is an absolute K such that W(r,k)>k^(k⌊log₂ r⌋/100000) for all k≥K and r≥2, so W(r,k)^(1/k) diverges for each fixed r≥2.

Curated contributions 1

Understand the problem ↓

Understanding the problem

Read the background

There is an absolute K such that W(r,k)>k^(k⌊log₂ r⌋/100000) for all k≥K and r≥2, so W(r,k)^(1/k) diverges for each fixed r≥2.

Why it matters

If correct, answers Erdős’s superexponential-growth question with a bound uniform in the number of colors.

What counts as progress

Do van der Waerden numbers grow superexponentially with progression length for every fixed number of colors?

Website timeline

Publication dates and editorial changes.