The authors had maintainers from scikit-learn, Sphinx, and pytest review 296 AI-generated PRs that passed SWE-bench’s automated grader and found roughly half would not be merged — maintainer merge rates are about 24 percentage points lower than the automated grader and show slower apparent improvement, suggesting benchmarks can overestimate real-world usefulness without human feedback or iteration.
Astrid Eichhorn is a leading proponent of asymptotic safety, a conservative quantum-gravity approach that envisions a fractal, scale-invariant spacetime at the Planck scale where couplings reach a fixed point. Her work emphasizes including matter–gravity interactions and linking Planck-scale behavior to lower-energy, testable physics.