P = NP might already be true, here's why

OpenAI has published a proof that the Exponential Time Hypothesis (ETH) is false. They used AI to create a subexponential algorithm for #3-SAT.

This implies P = NP is more likely, that there exists a polynomial-time algorithm (O(nᵐ)) for NP problems.

For over a year, I've been working on a concept of a polynomial-time SAT solver.

Algorithm concept

What ETH claimed

ETH said no algorithm could solve 3-SAT faster than O(2ⁿ). It was a conjecture, but a lot of work in the field rested on it. If you could solve 3-SAT fast, you could solve a lot of other NP problems fast. ETH said you couldn't.

OpenAI's proof breaks that. A subexponential algorithm for #3-SAT means the exponential lower bound people assumed doesn't hold.