Hot take: 🌶️ An agent should be generating nearly all code today
Hot take: 🌶️
An agent should be generating nearly all code today.
If the agent fails, the focus should be on improving instructions, prompts, architecture, tools, skills, tests, and feedback loops so it can reliably do so.
Typing code by hand should be a rare exception - for instance when I can clearly change the code more effectively than an agent.
If not, I’m behind, using outdated tools, and wasting my employer’s time and money.