Podcast Summary: The Daily – "Can A.I. Already Do Your Job?"
Date: February 18, 2026
Host: Natalie Kitroeff (with guest Kevin Roose, NYT tech columnist)
Theme: Exploring the recent breakthrough in “agentic” A.I. coding, and what it means for work, jobs, and the future of knowledge labor.
Episode Overview
The episode delves into one of the most transformative trends in A.I.: the rise of “agentic coding”, where A.I. models can autonomously write software, complete complex tasks, and even improve themselves—all with minimal human intervention. Natalie Kitroeff and Kevin Roose break down how these new tools work, their leap in capabilities, and the economic and social implications, notably the potential impact on jobs.
Key Discussion Points & Insights
1. What is "Vibe Coding" & How Did We Get Here?
- Definition:
- Vibe coding refers to using A.I. to write code on your behalf, without needing traditional programming experience.
- Coined by Andrej Karpathy, ex-OpenAI engineer, a year ago.
- Significance:
- “Instead of learning a programming language and writing the code line by line… let the program write the code for him.” – Kevin Roose (02:30)
- The real turning point: A.I. has advanced to the point where anyone can build software, not just coders.
- Early Limitations:
- “Honestly it was kind of clunky. The tools… you still needed to know a little bit about programming… pretty buggy.” – Kevin Roose (03:51)
- Initially, only truly useful to professional software developers.
2. The Shift: "Agentic Coding" and A.I. Teams
- Evolution:
- Agentic coding means A.I.s act as fully autonomous agents capable of managing whole projects independently.
- “With these new tools, you can just sort of give them a project and they can make a plan. They can decide what programming languages they want to use. They can even create their own little team of agents…” – Kevin Roose (04:20)
- Industry Impact:
- “They are now able to do the kinds of tasks that would take humans hours or sometimes even days to do. They are working their way up into the entry level functions in a lot of white collar industries.” – Kevin (05:07)
- A major leap in the models’ reasoning, accuracy, and autonomy.
3. The Claude Code Case Study (Anthropic’s Tool)
- Origin Story:
- Born as a side project by engineer Boris Czerny at Anthropic.
- Initially used by engineers, but rapidly adopted by non-technical staff (marketing, sales, finance) to automate parts of their jobs. (07:19–08:33)
- Viral Popularity:
- “Millions of people have been using it. They are doing increasingly complex tasks with them. …I don’t write any code anymore. I basically supervise this team of agents, and... go get lunch and I come back and the work is done.” – Kevin Roose (06:05–07:04)
4. Live Demonstration: Building a Website in Minutes (09:07–14:35)
- Process:
- Roose prompts Claude Code to make a sleek, professional, Eagles-colored personal website for Natalie—plus an embedded playable Tecmo Bowl-style football game.
- The A.I. writes 644 lines of code in just over a minute.
- “I feel pretty blown away. I honestly can’t believe that we just made my first website and that it has a video game on it.” – Natalie Kitroeff (14:41)
- Demonstrates how non-coders can direct agentic A.I. with simple instructions.
5. Broader Implications: Jobs, Talent, and the Economy
- Is This the A.I. Moment?
- “This feels different to me… when you can actually build something useful without knowing how to write code… this is what people think is the thing that’s going to… genuinely change how people work.” – Kevin (16:42)
- Evidence of Impact:
- Stanford research: employment for young software engineers is down 20% since 2022 as companies hire fewer coders, relying on A.I. tools. (18:31)
- A.I. as Workforce:
- “You could have a company with some human employees and then a whole bunch of A.I. agents doing tasks.” – Kevin (17:50)
6. How Good Are These Tools, Really?
- Verification & Self-Correction:
- “Coding is an interesting test ground... the code either runs or it doesn’t.” – Kevin (19:35)
- Not perfect yet; still needs human supervision, but improving rapidly.
- Self-Improving A.I.:
- “These A.I. tools are building versions of themselves. They are starting to improve themselves.” – Kevin (20:25)
- OpenAI’s GPT 5.3 Codex used itself to help train future versions. (20:53–21:26)
- “This is a sort of scenario that we’ve heard about in science fiction for years, …Recursive self improvement, building better and better versions of themselves.” – Kevin (21:35)
- Potential for “intelligence explosion”—still debated, but the acceleration is undeniable.
7. The Specter of Job Loss and Economic Disruption
- Industry Anxiety:
- “...AI could potentially eliminate half of all entry level white collar jobs within the next five years.” – Dario Amodei, CEO of Anthropic, quoted by Kevin (25:23)
- Even a partial realization would transform labor markets.
- “People are really uncertain. They don’t know if the skills they’re building are obsolete. College students don’t know what they should be studying… People are really getting nervous.” – Kevin (25:38–26:24)
- Will it Be as Big as Feared?
- Even the A.I. insiders admit there’s massive uncertainty on how quickly or extensively displacement will happen. (24:39–25:23)
8. The Uncertain Road Ahead
- What Comes Next?
- “A year from now these agentic tools will be dramatically better... They will become full-fledged members of the workforce. …a new kind of company [will emerge] with A.I. work at the center of it.” – Kevin (26:45)
- Both rapid improvement and unpredictability are signature features—Roose cautions he can’t predict beyond 6 months because of the breakneck pace. (27:30)
- Cautionary Note:
- “The most we can say right now is these tools are getting better at a very fast and accelerating rate, and... there’s a lot of unknowns about what that will mean for the future of human work.” – Kevin (27:59)
Notable Quotes & Timestamps
- “Let the vibes do the work.” – Kevin Roose (03:01)
- “You just have a computer that can use a computer.” – Kevin Roose (08:40)
- “My dream, a team of robots.” – Natalie Kitroeff (05:02)
- “I feel pretty blown away... It has a video game on it.” – Natalie Kitroeff (14:41)
- “AI could potentially eliminate half of all entry level white collar jobs within the next five years.” – Kevin Roose quoting Anthropic CEO Dario Amodei (25:23)
- “These AI tools are building versions of themselves. They are starting to improve themselves.” – Kevin Roose (20:25)
- “The most we can say right now is these tools are getting better at a very fast and accelerating rate…” – Kevin Roose (27:59)
Key Timestamps
- 01:12–03:17 – Vibe coding explained and early experiments
- 04:20–07:04 – Agentic coding & how these tools run “robot teams”
- 07:19–08:56 – Claude Code’s story and viral adoption
- 09:07–14:41 – Live demo: building Natalie’s website (w/ video game)
- 16:15–17:50 – The leap: Why agentic A.I. is a paradigm shift
- 18:31–19:14 – Signs of job displacement in tech jobs
- 20:25–21:35 – Self-improving A.I.; science fiction becomes real
- 24:39–25:23 – A.I. leaders’ warnings about job loss
- 26:45–28:08 – The next year: predictions, uncertainty, acceleration
Memorable Moments
- Natalie’s delighted awe as her personalized website (with a football game) is built on-air by Claude Code (14:04–14:41)
- The candid discussion of both excitement and anxiety from Kevin about the velocity of change and the lack of predictability (especially 25:38–27:30)
- The clear acknowledgment that for the first time, A.I. is not just a “fancy Google,” but doing real, valuable work autonomously (16:42–17:34)
Tone & Style
The episode is conversational, playful, and curious (plenty of banter between Natalie and Kevin), but also direct and earnest about the serious economic stakes and uncertainty posed by A.I. advances. The excitement of experimenting with agentic A.I. is tempered by real concern for what it means for employment and society.
Summary
In this episode, The Daily explores the startling leap of A.I. from simple chatbot to autonomous agent capable of “doing your job”—building real software, automating tasks, and even improving itself at a staggering pace. Host Natalie Kitroeff and tech columnist Kevin Roose offer a behind-the-scenes look at how these new tools work, their rapid evolution, and the implications—both exciting and unsettling—for the future of work. As A.I. edges closer to being a true participant in the workforce, listeners are left to ponder: How fast, and how far, will it go?