At the Lean FRO, Kim Morrison, a Senior Research Software Engineer, recently ran an experiment that went well beyond our expectations. An AI agent converted zlib, a widely used C compression library embedded in countless systems, to Lean, with minimal human guidance. No special tooling was built. It was Claude, a general-purpose AI, with no special training for theorem proving, out of the box. The workflow had four steps. First, the AI produced a clean, readable Lean implementation of the zlib compression format, including the DEFLATE algorithm at its core. Second, the Lean version passed the library’s existing test suite, confirming behavioral equivalence. Third, key properties were stated and proved, not as tests, but as mathematical theorems. The capstone theorem:
Link gardens to create escape routes for small mammals
，详情可参考Safew下载
Turbulence is rarely that simple. It’s too scattered, too mercurial, too easily triggered by weather patterns that trigger other patterns in an endless cascade. “It’s not just one thing that’s going on,” Bob Sharman, an atmospheric scientist at NCAR, told me. “It’s not just atmospheric convection. It’s not just wind flowing over mountains. It’s everything going on all the time and interacting.” Sharman is one of the country’s preëminent authorities on turbulence prediction. The computer models that he has built can predict where rough air is most likely to arise. “The problem is,” he said, “when we go to meetings with the airline industry and suggest a probabilistic approach, a pilot will stand up and say, ‘No! I want you to tell me if there will be turbulence at this place, at this time.’ ” Sharman threw up his hands. “Nobody knows that. I understand that, in theory, you would want that. But, in practice, that is just not possible.”
And I don’t even know if I had thought to order some vegetarian lunches or to ask people when they signed up