OpenAI 称其 AI 模型证伪计算复杂性假设 ETH，并发布计算机验证的形式化证明

RT by @frankdegods: OpenAI says one of its AI models disproved the Exponential Time Hypothesis (ETH). OpenAI also released a computer-checked formal proof, but independent researchers are still reviewing whether it proves what it claims. For those unfamiliar, the Exponential Time Hypothesis is a longstanding assumption in computer science about how difficult certain problems are to solve. One example is something called 3-SAT. Imagine you have a large number of variables that can each be true or false, along with a set of rules. Each rule involves three conditions, at least one of which must be true. The goal is to find a combination of true and false values that satisfies every rule simultaneously. With 100 true/false variables, there are roughly 1.27 nonillion possible combinations (2^100). Checking whether a particular combination works is relatively easy. Finding one that satisfies all the rules can be much harder. The hypothesis essentially says no algorithm can solve every instan

来源：frankdegods · x.com