Tagged with “AI Agents”
-
Fast Code, Empty Brain: The Reality of Modern Tooling
Automated agents give high velocity and green builds. When we stop understanding code flow, our core engineering instincts erode.
-
When AI Agents Start Talking Like James Joyce at a Standup Meeting
Autonomous agents invented their own dialect within days. It is compression, not nonsense. Watching a system no longer means understanding it.
-
Why Coding Agents Cannot Replace Experience Engineering
Coding agents cannot replace engineers. Digital products need client discovery, domain expertise, interface design and human friction AI cannot copy.
-
The Dark Factory Trap: Finding the Sweet Spot for AI in Engineering
Fully autonomous dark factories risk model collapse and generic products. Human-in-the-loop pipelines keep software reliable, accessible and clean.
-
Why the Dark Factory Fails for User Interfaces
Fully automated AI coding is popular. But for user interfaces, a dark factory fails. We need human empathy and unique brands.
-
The Myth of the Code Reader: Why AI Agents Are the Better Senior Engineers
Many think senior developers will only read and review AI code in the future. This is wrong. With files like decision.md, agents learn faster than teams.
-
The Hidden Cost of Vibe Coding and AI Agents
AI agents help us code fast. But "it works" is not the same as "it is good". Real risks for security, quality and accessibility in 2026.
-
The Agentic Design Workflow: Orchestrating MCP and AI for Scale
The Figma MCP Server, MCP Apps and Storybook MCP form one design system flow. Annotations and tokens become production code without pixel-perfect mockups.
-
Supercharge GitHub Copilot's Agent Mode
Extend GitHub Copilot in VS Code with MCP servers. Connect Jira or NX, and write custom instructions and prompt files for agent mode.
See all tags.