Quantum computers can solve problems significantly more efficiently than standard computers. One area where quantum computers are predicted to have a major impact is machine learning; several quantum algorithms were believed to be exponentially more efficient than their classical counterparts. However, a breakthrough result of Tang showed that many of these algorithms can be “dequantised”: there exist “quantum-inspired” classical algorithms based on the same ideas that are only polynomially slower. In this talk, I will introduce recent works on quantum-inspired classical algorithms for solving linear equations.