ADSP: Algorithms + Data Structures = Programs Episode 246: Not High on AI?
23 snips
Aug 8, 2025 The discussion dives into the effectiveness of AI-assisted coding tools, revealing a mix of enthusiasm and skepticism among developers. Trust in AI technologies is declining, influenced by personal experiences and environmental concerns. The hosts analyze productivity claims versus actual data, noting complexities in adoption across different experience levels. They also highlight the role of AI in generating boilerplate code, contrasting experiences with programming languages. Personal anecdotes underscore the value of mastering tools, enhancing overall programmer efficiency.
AI Snips
Chapters
Transcript
Episode notes
Developer Trust In AI Has Fallen
- Ben Deane and Conor cite the Stack Overflow survey showing developer trust in AI fell year‑over‑year.
- Early use has reduced average trust as developers discovered real-world limitations.
AI Code Reviews Can Break PRs
- Ben Deane describes AI code-review bots leaving so many comments a pull request became unreviewable.
- He spent 45 minutes vetting a dozen suggestions that were all wrong, killing the PR workflow.
AI's Hidden Environmental Cost
- Ben Deane highlights the large environmental footprint of AI and rising energy costs for simple queries.
- He warns the current AI path increases data-center energy demand dramatically.