Consultant: most companies fall into the skills/subagents trap, real agentic engineering is rare
alex_verem · x · 2026-11-16
In a retweeted thread, @mardehaym argues very few companies take agentic engineering seriously—most fall into the trap of piling up skills and subagents, because a new skill shows results the same afternoon while a test suite on every PR is boring and screenshot-unfriendly.
Evidence from practice: their managed delivery for a performance-marketing platform since 2023 shipped 22 production releases in Q1 2026 with zero rollbacks; the client's CTO ships code through his own AI agent through the same quality gates; when QA became the bottleneck, they reported it themselves.
Key point: when agents write most of the code, testing and review set the pace—the decade-old engineering practices matter more than new tricks.
More from coding & agent
- That vibe-coded internal tool now costs 10x what it saves, consultant warns — alex_verem · 2026-11-16
- MIRIX-2.0: Multi-Agent Memory System Hits High Accuracy Without Fine-Tuning — Wei-Chieh Huang · 2026-11-16
- EDNA ships 64kB behavioral engine with MCP server for continuous human verification — PossibilityQueasy949 · 2026-11-16
- Is Playwright becoming muscle memory? Dev argues lighter tools fit quick checks — Fresh_Temperature197 · 2026-11-16
- OpenSEO open-sources an Ahrefs alternative: self-hosted, pay-per-request, MCP for AI agents — thisdudelikesAI · 2026-11-16
- Unsloth Desktop lets local LLMs autonomously control a browser and complete real tasks — nevrekaraishwa2 · 2026-11-16