The argument that even the best public AI models will always lag far behind what labs can use internally

Writing about what they call AI math breakthroughs, a writer argues that even the best public models will always be far behind what labs can work with internally. They ask readers to consider what that gap means for intelligence-intensive industries.