Drew Breunig: Most agentic work will become workflows, not perpetual LLM calls
dbreunig · x · 2026-11-05
Drew Breunig (Agentic AI Foundation) argues that once an agent task repeats, it should be paved into a deterministic workflow rather than run on a large model. His example: a friend's agent used Opus 4.5 to check Strava rides, infer cycling pace, read the calendar, compute ferry timing, and announce when to leave home—genuinely useful, but essentially a code-shaped job, like "hiring a McKinsey consultant to mow your lawn." The model's real value is up front: scoping, designing, and iterating requirements. His method: find desire paths in agent usage, then pave them into workflows, freeing models for genuinely novel tasks.
Related event: Drew Breunig: Most Agent Work Will Solidify into Coded Workflows(3 posts)→
More from coding & agent
- Running Qwen 27B Q4 at 0.7 tok/s on a 4GB-VRAM laptop as a main coding model — Felix_455-788 · 2026-11-05
- His agent now auto-labels his Google Calendar in the background — steren · 2026-11-05
- The security startups racing to rein in AI agents — pstAsiatech · 2026-11-05
- OpenAI October dev roundup: Decisions API, 12-hour caching, 8x faster Codex tokens — OpenAIDevs · 2026-11-05
- Datadog previews Canvas, a new way to collaborate with AI agents on observability — farooqib · 2026-11-05
- One hour with Opus 5.5 fully fan-translates a 2004 GBA game, graphics included — ezshine · 2026-11-05