The Hidden Cognitive Cost of AI-Assisted Programming - Ragunath Jawahar

This talk examines the paradox of productivity in AI-assisted coding, revealing that experienced engineers can actually become 19% slower despite believing they are more efficient. The speaker discusses the cognitive overhead that emerges from relying on AI tools, shifting the focus from code creation to validation, review, and integration, which can lead to increased context switching. The session addresses how this rapid code generation can incur additional costs related to debugging and maintenance and offers strategies to mitigate these challenges.