Agentic coding turns engineering managers into caretakers of literal-minded LLMs
MarcJSchmidt · x · 2026-07-26
The post argues that managing agentic coding workflows is unlike managing human engineers: LLMs do not carry context reliably, do not learn from accountability, and can derail from a single bad spec line.
The author says the operator is really a caretaker for literal-minded systems that must be nudged through repeated sessions, rather than a classic engineering manager coordinating sane, self-correcting teammates.
More from coding & agent
- Built 47 Agents With n8n and Claude, Now Sharing 3 Prompts He Says Actually Work — Aiden_Tech_Ai · 2026-07-26
- A Sunday morning workflow: coffee, Codex, voice Jarvis, then build the cancer app — DeryaTR_ · 2026-07-26
- Red-team report says Memanto memory core had race conditions and fail-open bugs — Abject_Egg7543 · 2026-07-26
- opencode fixes local file attachments mislabelled as text/plain — wonderfulremark · 2026-07-26
- ChatGPT Sol helps a solo dev prototype five HTML game ideas for Steam Next Fest — 74BIT · 2026-07-26
- Built with Codex, a parameterized cat generator is now open source — ring_hyacinth · 2026-07-26