All Podcasts / The Daily / Can A.I. Already Do Your Job?
The Daily
The Daily

Can A.I. Already Do Your Job?

February 18, 2026 • 30m

Summary

⏱️ 7 min read

Overview

Kevin Roos explains the revolutionary development of 'agentic coding' - AI tools that can autonomously write software and perform complex tasks with minimal human oversight. Through a live demonstration of Claude Code, he shows how these tools have evolved from basic assistants to capable agents that are transforming the job market, particularly for entry-level programmers and white-collar workers.

What is Vibe Coding and How It Started

Kevin Roos introduces the concept of 'vibe coding,' a term coined by former OpenAI programmer Andre Carpathi about a year ago. Instead of manually writing code line by line, users could let AI programs build software for them by simply describing what they wanted. While the initial tools were clunky and still required some programming knowledge, they represented the first time non-coders could build functional software.

  • Vibe coding was coined by Andre Carpathi, a former OpenAI programmer, to describe using AI tools to write code instead of manually programming
  • Early vibe coding tools allowed people to build software without knowing how to code, though they were buggy and limited
  • Kevin built test apps like 'Lunchbox Buddy' that used photos to suggest lunch combinations, but the tools were mostly useful only to professional programmers
" Let the vibes do the work. Exactly. And suddenly, for the first time about a year ago, you didn't have to know how to code to build software. You could just use these tools. "

The Evolution to Agentic Coding

The major breakthrough came with 'agentic coding' - AI systems that can autonomously handle much more of the development process. These tools can create plans, choose programming languages, form teams of specialized agents, and work independently for hours. The improvement has been so dramatic that professional programmers at major AI companies report they no longer write code manually, instead supervising AI agents that complete tasks while they're away.

  • Agentic coding allows AI systems to work much more autonomously, creating plans and deciding what programming languages to use
  • These systems can create teams of specialized agents - one for research, one for building, one for testing
  • Models are getting better at reasoning, solving complex problems, and making fewer mistakes, surprising even their creators
  • Professional programmers at AI companies now supervise agent teams rather than writing code manually
" I don't write any code anymore. I basically supervise this team of agents and I kind of orchestrate them and set them off toward tasks. And then I go get lunch and I come back and the work is done. "

📚 5 more sections below

Sign up to unlock the complete summary with all insights, key points, and quotes

Listen on Apple Podcasts More Episodes from The Daily