The author’s follow‑up to a viral Hacker News post lays out why LLMs threaten his fintech role. He notes that large language models still stumble on niche regulations such as local tax codes, forcing legal teams to handle those details. Yet the same knowledge base he built now lives in ChatGPT Pro prompts, eroding his perceived advantage and raises concerns about career longevity.

He describes workarounds: generic design docs, extra state‑machine detail, and splitting sensitive implementation into multiple tickets. Adding E2E test tickets creates buffer to catch bugs and file improvement tickets before release. These tactics buy time but also signal a shift toward “AI‑native engineer” practices, where prompt engineering and adversarial model reviews supplement traditional coding and reduces reliance on senior reviewers.

The post warns that if model capabilities keep rising, software engineering could become a commoditized service, similar to copywriting’s collapse after ChatGPT. He argues that only a tiny elite will retain value, while the rest face redundancy across finance, law, and biology. The author concludes that the profession is already reshaped; adapting tools is now a survival requirement and risk losing talent quickly.