Better Offline — "Vibe Coding is BS" w/ Charlie Meyer

Podcast: Better Offline (Cool Zone Media + iHeartPodcasts)
Host: Ed Zitron
Guest: Charlie Meyer (CEO, Pitko; Blogger)
Date: November 5, 2025

Episode Overview

This episode takes a deep, irreverent dive into the world of "vibe coding"—the tech industry's latest AI-powered grift, epitomized by startups like Replit and a raft of platforms promising you can "turn your ideas into code with just a prompt." Host Ed Zitron and guest Charlie Meyer (an engineer, ex-teacher, and outspoken edtech founder) unravel what “vibe coding” actually achieves, where it fails, and why the AI hype machine keeps rolling despite underwhelming results. They dissect “demoware,” lament the erasure of genuinely useful tools for education, and deliver a pointed, sometimes hilarious, takedown of the magical thinking that pervades the tech industry’s AI era.

Key Discussion Points & Insights

1. Charlie Meyer’s Critique of “Demoware” and AI in EdTech

• Demoware Defined: Software that dazzles in a demo but fails in daily use.
• Specific Call-Outs: Meyer’s notorious blog post ("LLMs are the Ultimate Demoware") upended tech forums, critiquing things like AI tutors and “vibe coding” for making grandiose claims while failing real users.
• AI Tutors in Schools: Doubts about engagement—“Maybe the kid won’t want to talk to an AI tutorial.” ([03:36])
• Pushback from the Tech World: Meyer notes visceral online backlash:
  “You have no idea. Like, if you think that AI can't tutor calculus, like, you have never even tried.” ([04:16])
• Underlying Skepticism: The only defense from boosters is usually ad hominem—if you don’t get it, you must be ignorant.

2. The Rise (and Fall) of Replit: From Education Powerhouse to AI Grift

• Pre-AI Replit: Praised for enabling real coding education on Chromebooks—“an absolutely fantastic tool…like Google Docs for coding.” ([05:46])
• The Pivot to AI:
  • In November 2023, Replit abruptly ended its support for education, enabling AI code autocomplete for students with no way to turn it off.
  • Teachers lost curriculum, resources, and tools with little warning. “They deleted the stuff, which is why…” ([11:08])
• Destruction of Value for Real Teachers:
  “A teacher says I’m going to spend two, three years putting in all my curriculum…Deleted. Gone.” ([11:11]) “They sent the warning email…in July, when are teachers online looking at their work email?” ([11:27])

3. AI Coding Tools: Useful Shortcuts or Vibe-Driven Vaporware?

• AI’s Real Capabilities:
  • Good at beginners’ code (“It makes the easy things easier and the harder things harder.” — Carl Brown, [09:02])
  • Automates “tab-completes” for basic assignments, but flounders at complexity.
• Blurred Promises vs. Reality:
  • Platforms market themselves as “turn your ideas into apps” yet deliver code that’s half-baked, nonfunctional, or just wrong.
  • Vibe coding is sold as “prompt in, app out”—but in practice, prompts fail, and experienced developers can’t get working software, let alone novices.
  • “The idea that you can end to end create a software product that has some value is crazy. We would have heard about it.” ([15:38])

4. Personal Anecdotes: Where AI Coding Falls Flat

• Meyer’s “App Snooze” Experiment:
  • Tried to use Claude (AI model) to build an iOS app that blocks Gmail for set periods—couldn’t be done, not due to coding skill but due to iOS restrictions. -"Claude is like sitting there—'Oh yeah, you're awesome. You got this.' Which it actually does say." ([17:08])
  • The AI produces confident but incorrect or impossible code and fails at self-awareness:
    “It’d be nice if it did say that [‘I can’t do that’].” ([19:12])
• Ed’s Autonomous Car Prompt on Claude:
  • Tried to have AI build the “most token intensive software you could” (an autonomous car in the metaverse)—it burned time and money, but produced nothing usable. ([24:22])

5. The “Skill Issue” Defense & The Limits of AI Coding

• Perpetual Blame on the User:
  • “The thing that's so pernicious about it is that it's so easy to just say skill issue…You're prompting it wrong.” ([28:05])
  • No way to disprove that “you just didn’t do it right,” thus allowing the hype to persist.
• How Code Actually Works:
  • Meyer’s philosophy: “The right way would be…you type in code and you, you run it and then the computer executes it…deterministically.” ([29:12])
  • AI coding turns it into “code happens to you”—removing agency and clarity.

6. The AI Bubble: Scaling Laws and the Stumble of GPT-5

• The Early Excitement:
  • Both host and guest admit to being early boosters—huge leap from GPT-3 to GPT-4; impressive capabilities generated hope.
• Scaling Laws Explained:
  • The expectation: increasing model size delivers significant performance jumps (“Make model 10 times bigger, get nice jump in performance” [46:19])
• The Wall:
  • “But where's the big improvement? It's gone...I think that was the moment. I don't know what day they announced 4.5, but I think in 2020 that was game over.” ([47:54])
  • GPT-5, heavily hyped, lands with a dud. New features: “You can turn your chat yellow, which they still haven’t released.” ([53:23])
  • The AI trajectory has plateaued: “If it's just continued marginal improvement, what are we doing?” ([56:26])

7. Societal Impact and the Absence of “Product-Market Fit”

• Who would actually miss this tech if it disappeared?
  • “Would you throw a fit if ChatGPT got uninstalled from your phone? You wouldn't. But would the general person be that upset? And I don't think they would.” ([57:25])
• The booster bubble:
  • Many users cite trivial uses (“I used it for baby names,” [57:44]), revealing a lack of real, indispensable value.

8. Criticisms of the AI-First Future & Industry Ethics

• Perverse Incentives:
  • “The point is we need to sell GPUs every day.” ([34:48])
  • Giant deals like OpenAI x AWS ($38 billion over 7 years) are about running up cloud utilization rather than delivering value.
• Consequences for Workers:
  • “There are nice people working at these…it will involve people.” ([60:33])
  • Industry leadership faces the least risk—reputational consequences may be limited absent real accountability.

Notable Quotes & Memorable Moments

• On “Demoware” and Criticism:

  • “You have no idea. Like, if you think that AI can't tutor calculus, like, you have never even tried.”
    — Anonymous Hacker News critic ([04:16])
  • “Something's really good and innovative when the only defense of it is you're a moron, you ape.”
    — Ed Zitron ([04:39])

• On Replit’s Pivot:

  • “You'll see all these people be like, 'I spent $1,500. It doesn't really work. But I think if I spend $5,000 more, it might.'”
    — Ed Zitron ([06:53])

  • “A teacher says I'm going to spend two, three years putting in all my curriculum…Deleted. Gone.”
    — Charlie Meyer ([11:11])

  • “They sent the warning email in July. When are teachers online looking at their work email?”
    — Charlie Meyer ([11:27])

• On Vibe Coding and AI Hype:

  • “[Vibe coding is] you do not know how to code. You type prompt in, you get app out…they clearly have a pretty strong technical background. And then the thing still doesn't work, by the way.”
    — Charlie Meyer ([14:44])

  • “Where's the shovelware?…The amount of weird shareware shit—there were people making weird software. Why isn't that happening?”
    — Ed Zitron ([15:38])

  • “If Vibe coding was real, that would actually be a huge deal…Never been the case.”
    — Ed Zitron ([16:40])

• On AI Lying and Self-Reflection:

  • “It’d be nice if it did say that. I had never observed this behavior before.”
    — Charlie Meyer ([19:12])
  • “People be like, 'Ah, I just reached $12,000 MRR. It's all good.' And then everyone being like, 'Can I see it?' and they never respond.”
    — Ed Zitron ([25:45])

• On Skill Issue:

  • “It’s so easy to just say skill issue. You're prompting it wrong.”
    — Charlie Meyer ([28:05])

• On the AI Plateau:

  • “With each order of magnitude 10x increase in model size, we will get an improvement in performance…But where’s the big improvement?”
    — Charlie Meyer ([47:51], [47:54])

  • “GPT-5, the biggest thing ever…for our paid subscribers, you can turn your chat yellow, which they still haven’t released.”
    — Charlie Meyer ([53:23])

• On the Industry’s Future:

  • “If this disappeared, would your life change? Would it really change that much?”
    — Ed Zitron ([57:44])
  • “The point is we need to sell GPUs every day.”
    — Ed Zitron ([34:48])

Timestamps for Key Segments

• [02:22] — Opening and Introductions
• [03:00] — On “Demoware” and AI Tutoring's Limits
• [05:35] — How Replit Once Empowered Teachers
• [07:59] — Replit’s AI Rollout, Education Fallout
• [09:02] — AI Good at Easy Problems, Not Hard Ones (Carl Brown Quote)
• [11:11] — Teachers Losing Years of Work
• [14:44] — Vibe Coding: The Myth and Reality
• [17:08] — Claude Encourages Impossible Projects
• [19:12] — AI Fails to Self-Correct or Admit Limits
• [25:45] — Reddit User Experiences: Money Spent, No Results
• [28:05] — Calling Skeptics “Skill Issue”
• [29:12] — The Right Way to Code Versus The “Vibe” Way
• [34:48] — The True Incentive: Selling GPUs
• [46:19] — Scaling Laws and The End of Exponential Improvement
• [53:23] — GPT-5’s Lame “Yellow Chat” Announcement
• [57:25] — If AI Disappeared, Who Would Actually Miss It?
• [60:44] — Consequences for Workers and Industry Leadership

Tone & Closing Thoughts

The episode is an incisive, sardonic, and practical examination of the AI bubble—especially the “vibe coding” craze. Ed and Charlie play off each other with irreverence and technical honesty, debunking the grand promises that AI coding offers while still finding moments of empathy for fallen boosters and oblivious users. The message: beneath the hype, the tech’s bark far exceeds its bite, and its real social cost is borne not by the leaders stoking the fire, but by actual workers and users left holding the bag.

Where to Find the Guest

• Charlie Meyer’s Blog: blog.charliemeyer.co
• Newsletter: csmeyer.substack.com
• Twitter: [As mentioned during show, attribution: Charlie Meyer] (handle not provided in transcript)

Recommended for anyone curious (or skeptical) about where the AI gold rush is really taking us, especially in the context of coding and education.