TRE Python bindings show ReDoS resistance

After seeing antirez's case of adding TRE to Redis, Simon Willison took a closer look at Ville Laurikari's TRE. He used Claude Code to build an experimental Python binding based on ctypes.

Testing this binding with malicious regex patterns made TRE's strength against ReDoS clear. Compared to Python's standard re module, TRE was far more stable even with representative malicious patterns, and the key reason was its architecture, which does not support backtracking.

In benchmarks, TRE processed even 10 million character inputs quickly. As input size grew, execution time increased linearly, while re was relatively slow even on small inputs and showed the risk of exponential degradation.