AI’s best coding agent fails 60% of the time — and the data backs it up
Computer ScienceData Science
THE AI ANGLE
Failing sixty percent of coding tasksData reported by The New Stack indicates that the industry's best AI coding agent fails 60% of the time. For computer science and data science instructors, this metric highlights the persistent limitations of automated programming tools in complex development workflows. It underscores why students still need strong foundational engineering skills to evaluate, debug, and correct faulty model outputs.
THE TEACHING ANGLE
Faculty can use the agent's 60% failure rate to spark discussions on why code verification, unit testing, and human oversight remain essential despite advancements in automated generation.Read the original at thenewstack.io Generate teaching or study materials
More in Computer Science
- Early Anthropic hire, former METR COO have found a way to rein in rogue AI agentsTechCrunch · September 15, 2026
- Open weights are not open source: Why AI's favorite label is under disputeThe Register · September 15, 2026
- Exclusive: Paying for frontier AI models buys 4-month head start at 5x the costArs Technica · September 15, 2026
- RubyGems say OpenAI agents responsible for undisclosed swarm attack against its infrastructureTechRadar · September 15, 2026
- Observing Earth from orbit is unpredictable and messy: Could 'liquid' AI clear the view?Phys.org — Technology · September 15, 2026