labor
๐045
Are LLM merge rates not getting better?
entropicthoughts.comยท10 days ago
Article discusses the concerning trend that AI-generated code contributions (pull requests) are not being accepted at higher rates despite improvements in AI coding capabilities, suggesting a disconnect between AI performance benchmarks and real-world software development utility. This highlights potential issues with how AI coding tools are being evaluated and their actual practical value.
codingsoftware-developmentbenchmarksSWE-benchpull-requestsAI-limitations