 The paper presents a method for parallelizing quantum error correction, QEC, decoding to achieve almost arbitrary speed, removing the roadblock to scalability in leading approaches to QEC. This requires delaying some classical feed-forward decisions, but the slowdown is now only polynomial in the size of the QEC code, averting an exponential slowdown. The paper numerically demonstrates their parallel decoder for the surface code and shows no noticeable reduction in logical fidelity compared to previous decoders, while demonstrating the predicted speed-up. This article was authored by Lucas Gorick, Donnie Brown, Pinton M. Barnes, and others.