How to learn programming and CS in the AI hype era interview with dev and prof Mark Mahoney

Dr. Mark Mahoney, a former software developer and professor with 23 years of teaching experience, emphasizes the importance of learning programming through traditional, hands-on methods rather than relying on shortcuts. He highlights the value of platforms like Playback Press, which he used to provide free programming courses, and stresses the need for learners to develop problem-solving and debugging skills through trial and error. Regarding large language models (LLMs), Mahoney acknowledges their utility in non-technical tasks such as creating visualizations or classroom demonstrations but cautions against their use in complex software development due to risks of errors and misunderstandings. He advocates for iterative planning with LLMs, where learners refine their own plans before executing code, rather than relying solely on AI-generated solutions. Critical thinking and understanding the "why" behind code structure are prioritized over passive acceptance of LLM outputs.

Mahoney expresses concern that overreliance on LLMs could lead to shallow learning, undermining students ability to think independently or debug effectively. He warns of "tutorial hell," where learners avoid the struggle of mastering foundational concepts, and notes that financial and geographic disparities in LLM access may exacerbate educational inequities. While he sees LLMs as supplementary toolsparticularly for self-aware learners who actively engage with materialhe underscores the irreplaceable role of human instructors in providing personalized guidance, motivation, and contextual understanding. Traditional methods like textbooks and project-based learning, which foster resilience and deeper comprehension, remain essential. A hybrid approach combining LLMs with active, hands-on learning is recommended to balance efficiency with the development of durable skills.