Average number of hours between each #curl vulnerability report, since March 1 2026 (excluding July as we closed during that month): 16.2
During 2025 the average interval was 48 hours ...
Discussion
Average number of hours between each #curl vulnerability report, since March 1 2026 (excluding July as we closed during that month): 16.2
During 2025 the average interval was 48 hours ...
@bagder curious to know how AI slop rate was determined. Were the reports self identified as AI generated and then an assessment of quality made afterward?
@macprodigy I decide that in a 100% pure human-made decision in a system that certainly may be a bit flawed...
Is slop only the "utter bullshit" category or does it also include stuff like "this was documented not to be thread safe" if it's reported using LLM-generated text?
Updated #curl September numbers converted into slides for coming talks