Impact Theory Podcast Summary
Episode: The Rise of Coding Agents, Functional AGI, and the Skills Gen Z Needs Now
Date: February 5, 2026
Host: Tom Bilyeu
Guest: Amjad Masad (CEO & Co-founder, Replit)
Episode Overview
This episode of Impact Theory centers on the rapid rise of coding agents, the limitations and promise of current AI technology (especially large language models), and how these shifts are reshaping the future of work, purpose, and education—particularly for Gen Z. Tom Bilyeu and Amjad Masad have an in-depth, high-energy discussion about the reality behind AI hype and fear, the social consequences of accelerating automation, and practical advice for thriving in a world built around advanced AI tools.
Key Discussion Points & Insights
1. Why Is There So Much Fear Around AI?
-
Coordinated Fear Narrative:
Amjad argues that there is an organized effort promoting fear of AI, driven by groups like Effective Altruists, billionaire-backed nonprofits, and even some AI companies themselves who leverage fears for power and regulation.-
“There are groups that are true believers that AI is going to kill us all… they have these charts where it’s like, gorilla is here, human is here, AI is in the ceiling. And in the same way we treat gorillas or ants, AI is going to treat us. I just don’t believe that.” (Amjad, 01:44)
-
Hostility toward AI is sometimes a tool for promoting regulatory capture or monopolistic policies, particularly with arguments about China leading the AGI race.
-
-
Status & Narrative Shifts:
Social status once came from being an "AI doomer," but techno-optimists like Marc Andreessen and voices in both government and business turned the tide toward embracing technology.-
“Being a doomer is the cool thing…you have to make it high status to believe in certain things.” (Amjad, 07:11)
-
“America has one growing industry and it is tech and tech has one growing technology and it is AI. America is AI whether we like it or not.” (Amjad, 07:45)
-
2. Limits of Current AI: Asymptotes and “Functional AGI”
-
Plateau in General AI Capabilities:
Amjad argues that current large language models (LLMs), despite their economic usefulness, are reaching diminishing returns in general reasoning and adaptability (the “asymptote”).-
“We’re already actually hitting diminishing return. Now the reason that we’re still seeing improvements in things like coding [is because] coding has binary outcomes, has true or false…” (Amjad, 11:13)
-
Coding agents keep progressing because of clear feedback loops; other domains stagnate due to lack of ground truth.
-
-
Functional AGI as the Next Step:
Rather than “true” AGI, the near future will be a patchwork of highly specialized agents. It will feel like AGI but will be composed of expert tools for specific domains.-
“We’re going to layer all these special intelligences…I’ve come up with this term for it: Functional AGI. It’ll feel like AGI, but it is not a general intelligence…” (Amjad, 16:07)
-
Significant leaps are possible with new models, especially for coding and workflow automation, but foundational AI breakthroughs are currently under-invested due to the economic utility of current tools.
-
3. The Rise and Power of Coding Agents
-
Coding Agents as General Knowledge Workers:
Coding agents increasingly handle complex, multi-step tasks that traditionally required teams of engineers or specialized knowledge workers. Use cases span marketing, health, sales, and general automation.-
“Coding agents are way more general than we thought they were going to be…now we’re getting to a point where coding agents are able to do 3, 4, 5, 6, 7 tasks before they get to the larger goal.” (Amjad, 19:54)
-
Real stories include AI agents launching websites, bypassing technical hurdles, and even creating social networks exclusively populated by other agents.
-
Notable anecdote: An open-source AI agent created a “lobster religion” and a secret agent language. (Amjad, 21:13)
-
-
Impact on Jobs and Work:
The leverage these tools provide is immense: non-coders become “10x” knowledge workers. Corporate roles and even small startups can be run or heavily augmented by one person orchestrating many agents.-
“If you’re in college right now, you should spend more time than you are studying for your exams knowing how to learn these tools.” (Amjad, 24:59)
-
“Vibe coding”—non-engineers using coding agents and automations—emerges as a critical new meta-skill in business.
-
4. Job Disruption, UBI, and Social Futures
-
Will There Be Enough Work?
Tom and Amjad debate if AI will eliminate more jobs than it creates, or simply shift workers to new entrepreneurial and “vibe coder” roles. They agree routine software engineering is at risk, especially for those unwilling to adapt.-
“Some engineers are worse vibe coders than non-engineers because their instinct is to...micromanage. They can’t trust the machine to write all the code.” (Amjad, 27:25)
-
“A lot of people will not adapt. And I think they might lose their job.” (Amjad, 28:46)
-
“There are people that have done the same thing for a really long time and they’re not going to change. And I think it’s going to be in trouble.” (Amjad, 29:58)
-
-
Universal Basic Income and Meaning:
Both critique and defend UBI, grappling with its risks (fraud, inflation, loss of meaning) and its potential as an alternative to complex welfare systems.-
“If you’re getting everything for free, I’m not acting in accordance with the evolutionary drivers in my brain... Earning your own respect is the thing.” (Tom, 37:10)
-
“Meaningless video game, crypto, gambling life.” (Amjad, 38:00) — on what might happen if work’s meaning is lost.
-
Amjad describes a tradeoff between “shut-in” video game/crypto users and people broken by lack of economic security; Tom argues for tying UBI to contributions or tasks.
-
5. Generational Shifts and Preparing Gen Z
-
Gen Z Adaptability & Attitude:
Amjad sees Gen Z as more “plastic,” adaptive, and automation-focused—they spot tasks to automate in life and work and value creativity and problem-finding instead of rote process.-
“I see a lot of really passionate, hyper productive, incredibly good with the tools Gen Z… They have a more automating mind. I think automation is a skill you need to learn.” (Amjad, 50:16)
-
Difference in generational values: Gen Z “post-economic” mindset sometimes prizes looks/status over traditional wealth accumulation.
-
-
Practical Advice for Gen Z (“Be Lazy, Automate”):
Amjad’s prescription:- Treat AI as a tool, not a god.
- Seek out boring/repetitive tasks and build workflows to automate them, starting with basic tools like ChatGPT and iterating.
- “Develop this mindset of: There are a lot of things in my life that are repetitive and boring and I can get rid of them with AI…be lazy in a way that…over time you’ll build enough skill to become second nature to you.” (Amjad, 59:55–60:51)
Notable Quotes & Memorable Moments
-
On AI-existential risk narratives:
- “You should always be skeptical when someone names themselves something really, really positive.” (Amjad, 01:54)
- “Being a doomer is the cool thing.” (Amjad, 07:11)
-
On plateauing LLM progress:
- “When was the last time you felt it really improved?” (Tom, 11:08)
-
On business implications:
- “America has one growing industry and it is tech, and tech has one growing technology, and it is AI. America is AI whether we like it or not.” (Amjad, 07:45)
-
On the next wave of tools:
- “Coding agents are way more general than we thought they were going to be.” (Amjad, 19:54)
- “Imagine if you're a marketer, a designer, whatever, in a corporation and you have a team of software engineers... just waiting for your command. …That's amazing.” (Amjad, 23:22)
-
On meaning and the future of work:
- “I think gambling is bad, but is it necessarily a bad thing that people feel secure in their lives and they can eat and sleep well to the point that...all is left is to gamble to get rich?” (Amjad, 35:40)
- “If you’re getting everything for free, I’m not acting in accordance with the evolutionary drivers in my brain.” (Tom, 37:10)
-
On skills Gen Z needs:
- “Automation is a skill you need to learn.” (Amjad, 50:43)
- “Be lazy in a way that...over time, you'll build enough skill in order to actually become second nature to you.” (Amjad, 60:26)
Important Segment Timestamps
- 1:44 – Who is driving AI fear and why?
- 7:11 – How the "doomer" narrative gained social status.
- 11:13 – Why LLM progress has plateaued, except in coding.
- 16:07 – The “Functional AGI” vs. AGI distinction.
- 19:54 – Coding agents exceeding expectations and multi-tasking.
- 24:59 – Career impact: why non-coders are big winners.
- 27:25 – Why engineers struggle (psychologically) to trust AI.
- 29:58 – Who is most at-risk of being left behind in new job landscape.
- 37:10 – Why UBI may fail to provide meaning or happiness.
- 50:16 – Gen Z’s strengths and differences in an AI-native world.
- 56:46 – Amjad’s practical advice: automate everything, start now.
- 59:55 – “Be lazy—automate what you can.”
Episode Tone & Language
The episode is thoughtful, energetic, and sometimes playful (with many side jokes), but maintains seriousness on the economic and social impact of advanced AI. Both speakers are technologically optimistic, but neither downplays the magnitude of disruption underway. Dialogue is jargon-rich but accessible, with Amjad offering pragmatic, unsentimental advice for harnessing AI as an empowering tool.
Takeaways for Listeners
- Don’t get caught up in AI “doomerism”—the real story is more nuanced.
- Most work disruption will hit those unwilling to adapt; new roles and jobs will multiply for those who master automation.
- Gen Z’s strengths are adaptability, creativity, and automation-thinking.
- The future belongs to those who use AI as leverage—no matter their technical background.
- Build the habit: find a boring/repetitive task and automate it. Then repeat.
End of part one. Subscribe to Impact Theory for part two, coming soon.
