Better Offline – Hater Season: Corey Quinn
Podcast: Better Offline
Host: Ed Zitron (Cool Zone Media and iHeartPodcasts)
Guest: Corey Quinn (Chief Cloud Economist, Duckbill Group)
Release Date: March 4, 2026
Episode Overview
This episode kicks off "Hater Season" on Better Offline: host Ed Zitron is joined by cloud industry provocateur Corey Quinn for a no-holds-barred exploration of Amazon Web Services (AWS), the tech industry’s cloud boom, and the ongoing generative AI/capex gold rush. The conversation skewers AWS’s labyrinthine cost structure, questions tech’s AI megalomania, dissects Big Tech infrastructure spending, and challenges boosterish narratives around AI replacing jobs and revolutionizing software. The tone? Wry, irreverent, and very unafraid to call tech executives out.
Key Discussion Points & Insights
1. AWS Pricing Complexity & Why the Bill Keeps Going Up
Timestamp: [02:31] – [06:36]
- Per-usage Billing Confusion: AWS bills just keep growing, not due to secret price increases, but because of the messy nature of usage-based billing and human behavior.
- "The safe bias is: don’t turn things off... You’re not charged for the things you use in the cloud so much as you’re charged for the things you forget to turn off." (Corey Quinn, [06:14])
- Engineers Are Not Incentivized: Once infrastructure is spun up, there's little incentive to remove unused resources, leading to ‘cloud sprawl.’
- Is It a Conspiracy? Corey quips that AWS isn't competent enough to run such a scheme—“It is the nature of charging per usage. And honestly, assholes...” ([03:42])
2. Organizational Dysfunction at Amazon
Timestamp: [07:00] – [08:50]
- Too Many Services, Too Many Silos: Amazon is internally siloed with over 200 services; even employees can’t keep track.
- “I can speak incredibly convincingly about AWS services that do not exist to Amazon employees and not get called out on it...” (Corey Quinn, [07:29])
- Choosing the Right Service Is a Crapshoot: With 700+ VM options, even experienced engineers are guessing.
3. Amazon Leadership & The Technical Pressures of AI
Timestamp: [09:16] – [11:40]
- Andy Jassy’s Shift: Once a technical, hands-on leader at AWS, Jassy became “just a figurehead” as Amazon’s CEO—“When he gave keynote talks... you could see the irrepressible humanity leaking out around the edges. He was a person. Now he is a figurehead.” ([09:24])
- Cynicism Toward AI Hype: Zitron and Quinn skewer Jassy for misusing “generative AI” as a shiny buzzword for basic predictive tasks.
4. The Generative AI/GPUs Boom — Is Anyone Thinking Straight?
Timestamp: [12:34] – [20:09]
- Real Investment in Real Infrastructure: “They are real when it comes to this stuff. There are a number of other things... But when you see Ready to go...” (Corey Quinn, [13:14])
- Capex Explosion: Amazon, Alphabet, Microsoft, and Meta pour over $650B into infrastructure annually. Question: Is there real demand, or is it speculative FOMO?
- Is AI Spend Overhyped? “The AI spend in most of these companies hovers... between 5 and 7% of their total infrastructure spend.” (Corey Quinn, [15:13])
- Sustainability of Spending: Ziton questions: “What like 7 to 8% of a 300 million contract?... That’s dogshit... that’s not going to pay back that capex.” ([19:19])
5. AWS Getting Stuck in the AI Hypecycle
Timestamp: [20:15] – [24:31]
- Potential Overbuild: What if the AI bubble pops? Are Amazon’s new AI-focused data centers repurposable?
- “GPUs... we’re all going get really into online gaming in a few years. I don’t know.” (Corey Quinn, [20:12])
- Contract with OpenAI: Amazon's $38B deal with OpenAI is questioned—"Is OpenAI good for the money?" ([20:52])
- In-house Chips (Trainium/Inferentia): There's no visible real-world traction for Amazon’s custom AI chips.
6. ‘AI Psychosis’ & The Endgame of High-Stakes Cloud Building
Timestamp: [25:25] – [32:35]
- Are Cloud Providers Building for a Real Use Case?
- “What use cases today are we looking at where wow, if we had 10 times as much inference as we do today, then X would be possible. Solve for X. I’m not hearing it.” (Corey Quinn, [29:12])
- The AGI Mirage: The only plan seems to be “maybe we’ll invent AGI in 2029”—not much of a business case.
- The Myth of AI Replacing Work: Even boosters can’t clearly articulate what all this AI compute will accomplish—“I genuinely can’t get that answer out of anyone. Even the boosters.” (Ed Zitron, [29:23])
7. Amazon’s Future If the AI Hype Fizzles
Timestamp: [39:33] – [41:47]
- Will a Bubble Pop Cripple AWS? AWS’s core—nuts and bolts cloud—remains, but if contracts prove bad, systemic risk could hit the whole sector.
- “They have a banging business of providing computers to the world. The question is, what happens when there’s a giant hole in their balance sheet?” (Corey Quinn, [39:54])
- Government Bailouts? Both hosts mock the notion of a tech sector bailout.
8. Innovation Stagnation at AWS
Timestamp: [41:50] – [43:33]
- From Bold to Boring: AWS once shipped game-changing products (e.g. Lambda: serverless compute, [42:32]) but now just releases incremental, uninspired updates.
- “Now it’s not just you now it’s an entire sales team... There’s so much filler here. Where’s anything actually worth writing about?” (Corey Quinn, [41:50])
- New in LLMs: Amazon’s attempts at large language models (e.g. Nova) are lackluster compared to competitors—"They're racing neck and neck with Ed's Taxidermy and Frontier AI Lab." ([43:33])
9. Generative AI: Real Impact or Just Hype?
Timestamp: [45:06] – [48:04]
- AI's True Use Cases:
- Corey uses AI tools for curation and simple code generation in very narrow scenarios—point solutions, not revolution.
- “Is this worth changing the entire world for? No, no, it is not. But I'll make hay while the sun shines.” (Corey Quinn, [45:58])
- Translators and Stock Photographers: Only a few fields (e.g., translation, stock art) are seriously disrupted at present.
10. AI Won’t Replace Enterprise Software (SaaS): The 'Build It Yourself' Fallacy
Timestamp: [49:20] – [52:16]
- Replacing SaaS With Claude/AI is Overhyped: No, companies can’t just “Claude up” their own business applications overnight and expect robustness, compliance, and auditability.
- “You can easily generate a website that purports to do something. But what if they needed that thing to be correct?” (Corey Quinn, [49:47])
- Why Businesses Stick With SaaS: Not for code, but for support, compliance, audit trails, and reliability.
Notable Quotes & Memorable Moments
Ed Zitron ([01:58]):
"Now we're in hater season. Hater season is when I bring people on, allow them to get rude... which we've never referred to [tech execs] as fuckwits or morons or dumbasses or shitheads..."
Corey Quinn ([03:42]):
"I used to think it was a conspiracy, and then I realized Amazon is in no way, shape or form organizationally competent enough to pull something like that off."
Corey Quinn ([06:36]):
"You're not charged for the things you use in the cloud so much as you're charged for the things you forget to turn off."
Corey Quinn ([15:13]):
"The AI spend in most of these companies hovers... between 5 and 7% of their total infrastructure spend."
Ed Zitron ([29:38]):
"Are we going to summon God through JSON is step one and step two is we're going to ask God what to do."
Corey Quinn ([41:50]):
"It was first gradual, then all at once... there’s so much filler here. Where’s anything actually worth writing about?"
Corey Quinn ([45:58]):
"Is this worth changing the entire world for? No, no it is not. But I'll make hay while the sun shines."
Corey Quinn ([52:16]):
"This is not an AI phenomenon. When Dropbox launched, the number one comment on Hacker News was, 'This is just rsync with some extra bells and whistles.' This isn't a product."
Additional Highlights
- AWS’s ‘Compute Optimizer’ Tool: Corey is “thrilled” by Amazon’s AI-powered cost optimization tool, though he jokes its creators are probably “eyed for layoffs or PIPs” ([08:50]).
- AI as Executive Catnip: "AI psychosis" is discussed as a form of corporate leadership mania ([45:06]).
- Personal Use of AI: Corey’s favorite implementation of AI? An “AI-powered executive assistant” (Billy the Platypus) that sends tongue-in-cheek corporate rejections to spammy emails ([53:10]).
Conclusion
This episode provides a sardonic yet deeply informed look at how AWS—and Big Tech at large—have become ensnared in the cloud/AI hype cycle, risking billions on bets whose business cases are fuzzy at best. Corey Quinn brings gritty insight and humor to demystifying cloud economics, while Ed Zitron punctures techie self-delusion and AI snake oil. Listeners gain a much clearer picture of why cloud bills skyrocket, what’s real (and not) in AI, and where the limits of technology truly lie.
Suggested Segments:
- [02:31] – [06:36]: Why AWS bills never go down
- [15:13] – [20:09]: AI spending vs. actual business
- [29:12] – [32:35]: AI’s endgame and capex bubble
- [41:50] – [43:33]: AWS’s innovation slowdown
- [49:20] – [52:16]: Why AI can’t replace SaaS
For more irreverent tech analysis: