while the model companies all compete on the same benchmarks it seems likely their models will all converge towards similar outcomes unless something really unexpected happens in model space around those limit points…