In this episode of Dead Code, Jared speaks with software engineer Nicole Tietz-Sokolskaya about the ethical complexities of large language models (LLMs). Nicole shares her mixed experiences with LLMs—finding brief personal value but little long-term usefulness in her work—and critiques their environmental impact, reliance on questionable training data, and potential to concentrate power among major tech companies. She raises concerns about consent, bias, and the quality of generated code, while also acknowledging LLMs’ potential in accessibility and trust and safety roles. Ultimately, both Nicole and Jared advocate for a values-driven approach to technology, urging developers to critically evaluate the tools they use and the broader implications of their choices.
Links:
Mastodon: nicole@tietz.social
Dead Code Podcast Links:
Jared’s Links: