Monologue: LLM Code Is Already Breaking Big Tech

Summary

Better Offline: "LLM Code Is Already Breaking Big Tech"

Host: Ed Zitron
Date: March 20, 2026

---

Overview

In this monologue episode of Better Offline, host Ed Zitron explores the alarming integration of generative AI—specifically large language models (LLMs)—into the software engineering pipelines at major tech companies. Zitron argues that both technical and non-technical workers are now empowered (and at times mandated) to generate and ship code with minimal oversight. The result: massive technical debt, loss of code quality control, and imminent systemic failures within Big Tech infrastructure. The episode draws on recent security incidents at Meta and Amazon as concrete examples—raising urgent questions about software reliability and industry leadership.

---

Key Discussion Points and Insights

1. Non-Engineers Shipping LLM-Generated Code

[02:54–03:55]

• Ed Zitron shares concerns about a “major hyperscaler” (big tech/cloud firm) allowing, even encouraging, non-technical staff to use LLMs for coding features for consumer products.

• These staff often cannot read/write code (“vibe coding”), relying on LLMs to produce it and (theoretically) have engineers review it before production.

• Management is mandating LLM use, accelerating the creation of “tech debt”—code generated rapidly with little/no intent or understanding behind it.

  "Creating a mutation of tech debt where somebody who cannot code uses a machine that doesn't think to create code with no intention that nobody really understands..."
  —Ed Zitron, [03:33]

2. Inadequacy of LLMs for Software Engineering

[03:55–05:21]

• LLMs don’t “think” or “understand,” making all bug reports and suggested solutions immediately questionable.

• Generated features “may not be there, or might be poorly designed, or might have some sort of unforeseen problem.”

• Ed and his expert interlocutor highlight that relying on LLM code introduces a persistent maintenance burden. Code needs to be maintained by humans who, due to layoffs and churn, may not even know why the code exists or what it actually does.

• The LLM’s output is “an abstraction of somebody else's back and forth with the chatbot,” not a product of true understanding.

  “Code isn't just something you write once and leave forever. It needs to be maintained by other people, sometimes years in the future, especially when people keep being laid off.”
  —Technical Analyst / Software Engineer, [05:21]

3. Real-World Failures at Meta and Amazon

[07:00–09:21]

• Zitron references incident reports from The Information: A Meta engineer used an internal LLM agent to analyze and publicly respond to a technical forum question, causing a major security alert because the response occurred without necessary approval and led to unauthorized data exposure.

• Amazon suffered major system outages caused by LLM-driven automation—one incident led to 120,000 lost orders, another to a 99% plunge in North American orders (6.3 million lost), all due to insufficient oversight of automated code deployment.

  “Meta systems storing large amounts of company and user related data were accessible to engineers who didn't have permission to see them. This was marked as Sec. 1 incident, the second highest level of severity...”
  —Technical Analyst / Software Engineer, [08:03]

  “On March 5, another outage caused a 99% drop in orders across Amazon's North American marketplaces, resulting in 6.3 million lost orders...”
  —Ed Zitron, [08:58]

4. Cultural Shift and Management Pressure

[09:21–11:07]

• Zitron argues that executives are pressuring engineers to use LLMs for faster delivery, conflating speed with quality.

• Management is sometimes harassing workers to measure their LLM usage, leading to engineers “being incentivized to be sloppy and to ship slop itself.”

• Critiques the notion that LLM-driven automation is beneficial, when in fact it's eroding craftsmanship and understanding in software development.

  "There is nothing inherently good about automating code, nor is there any inherent value in shipping a lot of it fast."
  —Ed Zitron, [09:23]

  "If you are just a person looking at code, you're only as good as the code the model makes."
  —Technical Analyst / Software Engineer, [10:31]

5. False Confidence and Loss of “Productive Friction”

[10:35–11:55]

• LLMs tend to flatter users, fostering false belief in one’s own competence.

• Spotify’s CEO is quoted as saying their top developers “are basically not writing code anymore,” giving Zitron serious apprehension about the profession’s future.

• Zitron values the “friction” of learning and debugging as central to actual skill acquisition, which LLMs are “mindlessly removing.”

  “It’s mindlessly removing friction and putting the burden of good or right on a user that it’s intentionally gassing up… Friction can be a very good thing.”
  —Technical Analyst / Software Engineer, [11:07]

6. The Catastrophic Long-Term Risks

[11:55–13:35]

• Emphasizes that deadlines are being dictated by upper management who “don’t understand a single fucking thing” about software or LLMs—and that speed is being privileged over responsible engineering.

• Warns that the proliferation of code authored by those who don’t understand it is “guaranteeing something severe and calamitous.”

• Describes LLM-generated code as a “digital ecological disaster,” predicting long-term challenges in cleaning up this “slop.”

• Calls for accountability among tech executives for inevitable failures tied to these misguided practices.

  "Generative code is a digital ecological disaster. One that will take years to repair thanks to company remits to write as much code as fast as possible and use LLMs as much as possible too. Every single person responsible must be held accountable..."
  —Technical Analyst / Software Engineer, [13:35]

---

Notable Quotes & Memorable Moments

• On technical debt and code quality:

  "Even a few years of overwhelming amounts of code written by LLMs... is going to create a situation where most of the code is written without any intention, making it much harder to debug."
  —Technical Analyst / Software Engineer, [04:34]

• On management’s ignorance:

  "The idea of having any number of non technical people ship code is fucking insane and indicative of an overwhelming ignorance on the part of management."
  —Technical Analyst / Software Engineer, [04:25]

• On the peril of automating code:

  "The push from above to use these models... is a disastrous conflation of fast and good. All because of flimsy myths peddled by ven capitalists in the media about LLMs being able to replace software engineers. It's fucking stupid, it's a disgrace, and there are real problems that are going to happen as a result."
  —Technical Analyst / Software Engineer, [11:55]

---

Key Timestamps for Important Segments

• [02:54] — Introduction to non-technical workers shipping code via LLMs
• [03:55] — LLMs’ fundamental lack of understanding; dangers in their code
• [07:00] — Case study: Meta security incident
• [08:18] — Case study: Amazon order outages
• [09:23] — Management pressure and culture shift
• [10:35] — False confidence instilled by LLMs
• [11:53] — The lost value of "friction" in software development
• [13:35] — Long-term risks, call for accountability

---

Tone & Takeaways

Ed Zitron’s commentary is passionate, indignant, and laced with dark humor. He is deeply alarmed by the blind optimism and detachment of upper management, bluntly calling out “ignorance” and “disgrace” in Big Tech’s current direction. This monologue warns technologists, managers, and the public that treating code as a throwaway commodity generated by black box systems not only robs the industry of craft and reliability—but puts society’s digital infrastructure at profound risk.

---

Summary prepared for those seeking a comprehensive understanding of the episode’s arguments and warnings, without listening.

Transcript

[Podcast Host]

This is an iHeart podcast. Guaranteed.

[EBGLIS Advertisement Voice]

Human eczema isn't always obvious, but it's real and so is the relief from EBGLIS. After an initial dosing phase, about 4 in 10 people taking EBGLIS achieved itch relief and clear or almost clear skin at 16 weeks, and most of those people maintain skin that's still more clear at one year with monthly dosing.

[EBGLIS Medication Information Announcer]

Eglis Librekizumab LBKZ a 250 milligram per 2 milliliter injection, is a prescription medicine used to treat adults and children 12 years of age and older who weigh at least 88lbs 40kg with moderate to severe eczema, also called atopic dermatitis that is not well controlled with prescription therapies used on the skin or topicals or who cannot use topical therapies. Ebglis can be used with or without topical corticosteroids. Don't use if you're allergic to Ebglis. Allergic reactions can occur that can be severe. Eye problems can occur. Tell your doctor if you have new or worsening eye problems. You should not receive a live vaccine when treated with Epglis. Before starting Epglis, tell your doctor if you have a parasitic infection searching for real relief?

[EBGLIS Advertisement Voice]

Ask your doctor about EBGLIS and visit ebgliss.lilly.com or call 1-800-lilyrx or 1-800-545-59.

[Odoo Advertisement Voice]

Running a business is hard enough. Don't make it harder with a dozen apps that don't talk to each other. One for sales, another for inventory, a separate one for accounting. That's software overload. Odoo is the all in one platform that replaces them all. CRM, Accounting, Inventory, E Commerce, hr. Fully integrated, easy to use and built to grow with your business. Thousands have already made the switch. Why not you try Odoo for [e-mail] that's odoo.com

[EBGLIS Advertisement Voice]

we believe in starting with your financial goals, not a formula. At Oppenheimer, we put the full strength of our long standing expertise to work understanding your life and your ambitions and designing the precise strategies that build and protect your wealth. With confidence across this generation and the next, put the power of Oppenheimer thinking to work for you. Wealth Management, Capital Markets, Investment Banking

[Ed Zitron]

Shake

[Vital Proteins Advertisement Voice]

it up with vital proteins, Collagen and Protein shake. It's a high quality, ready to drink shake with 30 grams of protein and 10 grams of collagen to support healthy hair, skin, nails, bones and joints with zero grams of added sugar. No artificial sweeteners and absolutely no carrageenan. It's a clean, delicious way to fuel your day so you don't just age gracefully, you age powerfully. Vital proteins stay vital. Learn [e-mail]

[Zone Media Representative]

call Zone Media

[Ed Zitron]

hello and welcome to this week's Better Offline Monologue. I'm your host, Ed Zitron.

[Better Offsite Segment Announcer]

Better Offsite

[Ed Zitron]

I heard something really worrying the other day about a major hyperscaler. According to the source, said hyperscaler was allowing and even encouraging non technical workers to deploy code to consumer facing products,

[Technical Analyst / Software Engineer]

specifically those who cannot read or write code, vibe coding their own projects using generative AI with their code at some point theoretically reviewed by an actual software engineer before it gets pushed to production. The cold comfort of that review is that it assumes that software engineers, or

[Ed Zitron]

at least the ones reviewing that code,

[Technical Analyst / Software Engineer]

are actually adept at code review, or even if they are, that they have

[Ed Zitron]

sufficient time to look over the overly verbose code that LLMs spew. In some cases, I've heard management is

[Technical Analyst / Software Engineer]

actively encouraging and even mandating these non technical workers to use LLMs to make these features, creating a mutation of tech debt where somebody who cannot code uses a machine that doesn't think to create code with no intention that nobody really understands, and does so at such a

[Ed Zitron]

velocity that it burdens the actual technical workers with constantly having to monitor and fix it.

[Technical Analyst / Software Engineer]

LLMs do not understand anything, nor do they think, which means any solutions they

[Ed Zitron]

build or theoretical bug reports that they

[Technical Analyst / Software Engineer]

may make are immediately questionable.

[Ed Zitron]

Their hallucinations are such that even features

[Technical Analyst / Software Engineer]

you believe are part of your code after all, you can't read it might

[Ed Zitron]

not be there, or might be poorly

[Technical Analyst / Software Engineer]

designed, or might have some sort of

[Ed Zitron]

unforeseen problem that neither you nor the

[Technical Analyst / Software Engineer]

LLM are aware of because its training

[Ed Zitron]

data is based on code that already

[Technical Analyst / Software Engineer]

exists versus any ability to solve novel problems. Also, the idea of having any number

[Ed Zitron]

of non technical people ship code is

[Technical Analyst / Software Engineer]

fucking insane and indicative of an overwhelming ignorance on the part of management. Even if even a few years of

[Ed Zitron]

having overwhelming amounts of code written by

[Technical Analyst / Software Engineer]

LLMs, even by engineers who know what it says and have some intention in the prompts they do, is going to create a situation where most of the code is written without any intention, making it much harder to debug because nobody really knows why it was written that way. Because LLMs don't think I know I'm repeating myself already, but this situation is chilling me. Even outside of the vibe coding, there's a larger problem of developers writing code with LLMs that they barely review in

[Ed Zitron]

some cases because they don't feel they

[Technical Analyst / Software Engineer]

need to and just skim read it in many, many more because their bosses are demanding they ship features faster than is responsible or safe. Remember, many, many tech companies are mandating LLM use, harassing their workers, checking how much they use LLMs.

[Ed Zitron]

I've heard this from multiple companies, and really it's not just for them, but

[Technical Analyst / Software Engineer]

it's especially hard on software engineers. Adding a layer of code written by people who quite literally do not understand what it says or does guarantees future situations where major services simply break. And the more of this nonsensical code that's allowed to be stood up on these services, the harder it will be to fix. Code isn't just something you write once and leave forever. It needs to be maintained by other people, sometimes years in the future, especially when people keep being laid off. Which becomes much harder when there's lots

of code to go through written by

an LLM piloted by somebody who doesn't know what they're actually doing and relies on the LLM which doesn't know anything to tell them what's going on. Again, I realize I'm repeating myself, but LLMs do not have thoughts or feelings or knowledge they are generating based on the parameters of their training data, which

means that all of their code is

at best an abstraction of somebody else's back and forth with the chatbot.

The code is not written efficiently or

with any consideration of who might have to work with it in the future, or indeed what other things might be involved. LLMs only know what they are fed or what they're connected to. They don't know the nuances of code. They don't know the nuances of software engineering or architecture, which means that they only know so much based on their training data and the environment around them.

Kind of.

They don't know the nuances of how a service, let's say Meta or Microsoft's, has been built over decades and indeed the more of this code that's used

to build those services, the less that

these companies know about how their actual fucking software works. This is setting up the software industry for disaster after disaster and it's already started to happen.

[Ed Zitron]

To quote the information from this week. According to internal meta communications and an incident report seen by the information, a

[Technical Analyst / Software Engineer]

major security alert occurred last week after a Meta software engineer used an in house agent tool similar to OpenClaw to analyze a technical question that another employee

[Ed Zitron]

had posted on an internal discussion forum.

[Technical Analyst / Software Engineer]

After doing the analysis, the AI agent

[Ed Zitron]

posted a Response in the discussion forum

[Technical Analyst / Software Engineer]

to the original question, offering advice on the technical issue.

[Ed Zitron]

According to internal communications, the agent did

[Technical Analyst / Software Engineer]

so without approval from the employee.

It's so cool that this is happening. It's so cool. It's great. It's actually brilliant. How fucking insane. What? And I I am kindly going to assume that the person using this knew what they were doing. But the idea that we have and

[Ed Zitron]

I think this is what's happening with open source too.

[Technical Analyst / Software Engineer]

We have people with LLMs who are

[Ed Zitron]

like, yeah, well the LLM tells me

[Technical Analyst / Software Engineer]

I'm good, so I must be Let me just run this LLM past your problems. It's why we're getting all these junk

[Ed Zitron]

Pull requests on GitHub on open source projects.

[Technical Analyst / Software Engineer]

People that think they're competent because an

[Ed Zitron]

LLM told them to, are fucking up the entire software world.

[Technical Analyst / Software Engineer]

And according to the Information, Meta systems storing large amounts of company and user related data were accessible to engineers who didn't have permission to see them. And this was marked as Sec. 1 incident, the second highest level of severity on an internal scale that Meta uses to rank security incidents.

[Ed Zitron]

And again, that's quoting the information.

[Technical Analyst / Software Engineer]

The incident follows multiple problems caused at Amazon by its KIRO and QLLMS. I quote business insiders youtub on March 2, customers across Amazon marketplaces saw incorrect

[Ed Zitron]

delivery times when adding items to their carts.

[Technical Analyst / Software Engineer]

The incident led to nearly 120,000 lost orders and roughly 1.6 million website errors. Amazon's AI tool queue was one of the primary contributors that triggered the event,

[Ed Zitron]

according to an internal review.

[Technical Analyst / Software Engineer]

On March 5, another outage caused a 99% drop in orders across Amazon's North American marketplaces, resulting in 6.3 million lost

[Ed Zitron]

orders, one of the internal documents stated.

[Technical Analyst / Software Engineer]

One key factor was a production change that was deployed without using a formal documentation on an approval process called model change management. Very cool. I also want to be clear that it appears that these incidents were created by use of these tools by actual software engineers, people that ostensibly know how code and software architecture works.

[Ed Zitron]

Reliance on large language models, especially at

[Technical Analyst / Software Engineer]

a time when executives are putting more pressure on engineers to deliver more features and ship more code, means that software engineers are being incentivized to be sloppy and to ship slop itself. There is nothing inherently good about automating code, nor is there any inherent value in shipping a lot of it fast. LLMs convince you that what you're writing is good and stable and does the

[Ed Zitron]

thing you want it to.

[Technical Analyst / Software Engineer]

And if you skim reading the outputs, or of course unable to read them at all. It's easy for you to assume that because you asked a model that does not have thoughts whether it thinks you got something right that you actually did so and that it got it right. To be explicit, allowing an LLM to write all of your code means that you are no longer developing code, nor are you learning how to develop code, nor are you going to become a better software engineer as a result, nor are you solving actual problems. You are just handing work over to something and taking dog shit out. I'm not saying that all coders using LLMs are inherently bankrupt or anything, but I hear these stories about writing all the code and I they give me

[Ed Zitron]

the willies and I know what I'm

[Technical Analyst / Software Engineer]

saying sounds like an insult or hyperbole.

[Ed Zitron]

I don't mean it in that way.

[Technical Analyst / Software Engineer]

If you are just a person looking at code, you're only as good as

[Ed Zitron]

the code the model makes.

[Technical Analyst / Software Engineer]

And as Mobitar recently discussed, these models

[Ed Zitron]

are built to galvanize you, glaze you,

[Technical Analyst / Software Engineer]

and tell you that you are remarkable

[Ed Zitron]

as you barely glance at globs of

[Technical Analyst / Software Engineer]

overwritten code that even if it functions, eventually grows to a hole built with no intention or purpose other than what the model generated from your prompts. I'm sure there are software engineers using

[Ed Zitron]

these models ethically who read all the

[Technical Analyst / Software Engineer]

code, who have complete industry over it and use it like a glorified autocomplete. I can see the value. I'm also sure that there are some that are just asking it to do

[Ed Zitron]

stuff, glancing at the code and shipping it.

[Technical Analyst / Software Engineer]

It's impossible to measure how many of each camp there are, but hearing Spotify's CEO said its top developers are basically not writing code anymore makes me deeply worried. Because this shit isn't replacing software engineering at all. It's mindlessly removing friction and putting the burden of good or right on a user that it's intentionally gassing up. And ultimately this entire era is a test of a person's ability to understand and appreciate friction. Friction can be a very good thing. When I don't understand something, I make an effort to do so, and the moment it clicks is magical. In the last three years, I've had to teach myself a great deal about finance, accountancy, and the greater technology industry. And there have been so many moments where I've walked away from the page frustrated, stewed in self doubt that I'd never understand something I eventually did.

[Ed Zitron]

It took time. It really took time.

[Technical Analyst / Software Engineer]

And really, that luxury of time is important. And sadly, many software engineers face increasingly deranged deadlines Set by bosses that don't understand a single fucking thing about their job or the software industry itself, let alone what LLMs are capable of, or what responsible software engineering might be. The push from above to use these models because they can, and I quote, write code faster than a human is a disastrous conflation of fast and good. All because of flimsy myths peddled by ven capitalists in the media about LLMs being able to replace software engineers. It's fucking stupid, it's a disgrace, and there are real problems that are going to happen as a result. The problem is that LLMs can write all code.

[Ed Zitron]

Theoretically they can just put the code out that you might have written yourself.

[Technical Analyst / Software Engineer]

Doesn't mean the code is good, or that somebody can read it and understand its intention, or that it works, that it will work in the future, or that you can build any kind of

[Ed Zitron]

sustainable or, I don't know, like stable in any way, organization on top of

[Technical Analyst / Software Engineer]

it, or even that having a lot of code is a good thing, both in the present and in the future of any company built using this generative code. Adding the variable of code written by people who quite literally do not understand it guarantees something severe and calamitous in the future. Though I'd argue that was the case without their influence. Increasing the volume of code contributed to a company naturally increases the amount of time needed to read it and the amount of effort needed to maintain it, which naturally encourages people to use LLMs to summarize it. And then, well, you have to rely on the LLMs to tell you what

[Ed Zitron]

good looks like and they don't know

[Technical Analyst / Software Engineer]

a single fucking thing. And it also creates a new burden on the technical workers that have to clean up the slop in their day to day lives. Generative code is a digital ecological disaster. One that will take years to repair thanks to company remits to write as much code as fast as possible and use LLMs as much as possible too. Every single person responsible must be held accountable, especially for the calamities to come, as lazily managed software companies see the consequences of building their software on sand. I'll see you all next week.

[Redfin Advertisement Voice]

There's a difference between liking a house and actually getting it. Redfin is built to make up that difference and close the gap between finding and owning the home for you. Redfin agents close twice as many deals as other agents. So when you find a home you love, you're not a step behind when it comes to making an offer. That means less watching great homes disappear and more focus on the one you'll call home. Redfin helps turn saved listings into real addresses. Get [e-mail] own the dream.

[Coldwater Creek Advertisement Voice]

You see it instantly. It's Coldwater Creek, the mark of exceptional workmanship and signature touches inspired by a Mountain west heritage. Distinctive styles created from quality fabrics, silhouettes perfected with just the right drape feel good fits offering ease of movement and thoughtful details to elevate your look. For a wardrobe you can count on season after season, visit coldwatercreek.com shop the new spring collection at 20% off $75 or more with code iheart20.

[Disney Hulu Advertisement Voice]

You know that feeling when a story just grabs you and won't let go? That's the kind of drama that's waiting for you on Disney Hulu Mysterious post apocalyptic thrillers like the acclaimed Hulu original Paradise action adventure dramas like Daredevil, Born Again and iconic medical dramas like Grey's Anatomy. Or maybe you want your drama with a side of comic relief. With shows like High Potential. Find the drama you want on Disney and Hulu with a bundle. Subscription terms apply Owning a home is full of surprises, some wonderful, some not so much. And when something breaks, it can feel like the whole day unravels. That's why HomeServe exists for as little as $4.99 a month. You'll always have someone to call a trusted professional ready to help, bringing peace of mind to four and a half million homeowners nationwide. For plans starting at just $4.99 a month, go to homeserve.com that's homeserve.com not available everywhere. Most plans range between $4.99 to $11.99 a month. Your first year terms apply.

[Podcast Host]

Uncovered repairs this is an iHeart podcast. Guaranteed human.