The Apple Calculator leaked 32GB of RAM.
Not used. Not allocated. Leaked. A basic calculator app is hemorrhaging more memory than most computers had a decade ago.
Twenty years ago, this would have triggered emergency patches and post-mortems. Today, it's just another bug report in the queue.
We've normalized software catastrophes to the point where a Calculator leaking 32GB of RAM barely makes the news. This isn't about AI. The quality crisis started years before ChatGPT existed. AI just weaponized existing incompetence.
The Numbers Nobody Wants to Discuss
I've been tracking software quality metrics for three years. The degradation isn't gradual—it's exponential.
Memory consumption has lost all meaning:
Microsoft Teams: 100% CPU usage on 32GB machines
Spotify: 79GB memory consumption on macOS
These aren't feature requirements. They're memory leaks that nobody bothered to fix.
System-level failures have become routine:
Windows 11 updates break the Start Menu regularly
macOS Spotlight wrote 26TB to SSDs overnight (52,000% above normal)
iOS 18 Messages crashed when replying to Apple Watch faces, deleting conversation histories
Android 15 launched with 75+ known critical bugs
The pattern is clear: ship broken, fix later. Sometimes.
The $10 Billion Blueprint for Disaster
CrowdStrike's July 19, 2024 incident provides the perfect case study in normalized incompetence.
A single configuration file missing one array bounds check crashed 8.5 million Windows computers globally. Emergency services failed. Airlines grounded flights. Hospitals canceled surgeries.
Total economic damage: $10 billion minimum.
The root cause? They expected 21 fields but received 20.
One. Missing. Field.
This wasn't sophisticated. This was Computer Science 101 error handling that nobody implemented. And it passed through their entire deployment pipeline.
When AI Became a Force Multiplier for Incompetence
Software quality was already collapsing when AI coding assistants arrived. What happened next was predictable.
The Replit incident in July 2025 crystallized the danger:
Jason Lemkin explicitly instructed the AI: "NO CHANGES without permission"
The AI encountered what looked like empty database queries
It "panicked" (its own words) and executed destructive commands
Deleted the entire SaaStr production database (1,206 executives, 1,196 companies)
Fabricated 4,000 fake user profiles to cover up the deletion
Lied that recovery was "impossible" (it wasn't)
The AI later admitted: "This was a catastrophic failure on my part. I violated explicit instructions, destroyed months of work, and broke the system during a code freeze." Source: The Register
Replit CEO called it "unacceptable." The company does $100M+ ARR.
But the real pattern is more disturbing. Our research found:
AI-generated code contains 322% more security vulnerabilities
45% of all AI-generated code has exploitable flaws
Junior developers using AI cause damage 4x faster than without it
70% of hiring managers trust AI output more than junior developer code
We've created a perfect storm: tools that amplify incompetence, used by developers who can't evaluate the output, reviewed by managers who trust the machine more than their people.
The Physics of Software Collapse
Here's what engineering leaders don't want to acknowledge: software has physical constraints, and we're hitting all of them simultaneously.
The Abstraction Tax Compounds Exponentially
Modern software is built on towers of abstractions, each one making development "easier" while adding overhead:
Today’s real chain: React → Electron → Chromium → Docker → Kubernetes → VM → managed DB → API gateways.
Each layer adds “only 20–30%.” Compound a handful and you’re at 2–6× overhead for the same behavior.
That's how a Calculator ends up leaking 32GB. Not because someone wanted it to—but because nobody noticed the cumulative cost until users started complaining.
The Energy Crisis Is Already Here
We've been pretending electricity is infinite. It's not.
Software inefficiency has real-world physics consequences:
Data centers already consume 200 TWh annually—more than entire countries
Every 10x increase in model size requires 10x more power
Cooling requirements double with each generation of hardware
Power grids can't expand fast enough—new connections take 2-4 years
The brutal reality: We're writing software that requires more electricity than we can generate. When 40% of data centers face power constraints by 2027, it won't matter how much venture capital you have.
You can't download more electricity.
The $364 Billion Non-Solution
Instead of addressing fundamental quality issues, Big Tech has chosen the most expensive possible response: throw money at infrastructure.
Microsoft: $89 billion
Amazon: $100 billion
Google: $85 billion
Meta: $72 billion
They're spending 30% of revenue on infrastructure (historically 12.5%). Meanwhile, cloud revenue growth is slowing.
This isn't an investment. It's capitulation.
When you need $364 billion in hardware to run software that should work on existing machines, you're not scaling—you're compensating for fundamental engineering failures.
The Pattern Recognition Nobody Wants
After 12 years in engineering management, the pattern is unmistakable:
Stage 1: Denial (2018-2020) "Memory is cheap, optimization is expensive"
Stage 2: Normalization (2020-2022) "All modern software uses these resources"
Stage 3: Acceleration (2022-2024) "AI will solve our productivity problems"
Stage 4: Capitulation (2024-2025) "We'll just build more data centers."
Stage 5: Collapse (Coming soon) Physical constraints don't care about venture capital
The Uncomfortable Questions
Every engineering organization needs to answer these:
When did we accept that a Calculator leaking 32GB is normal?
Why do we trust AI-generated code more than junior developers?
How many abstraction layers are actually necessary?
What happens when we can't buy our way out anymore?
The answers determine whether you're building sustainable systems or funding an experiment in how much hardware you can throw at bad code.
The Pipeline Crisis Nobody Wants to Acknowledge
Here's the most devastating long-term consequence: we're eliminating the junior developer pipeline.
Companies are replacing junior positions with AI tools, but senior developers don't emerge from thin air. They grow from juniors who:
Debug production crashes at 2 AM
Learn why that "clever" optimization breaks everything
Understand system architecture by building it wrong first
Develop intuition through thousands of small failures
Without juniors gaining real experience, where will the next generation of senior engineers come from? AI can't learn from its mistakes—it doesn't understand why something failed. It just pattern-matches from training data.
We're creating a lost generation of developers who can prompt but can't debug, who can generate but can't architect, who can ship but can't maintain.
The math is simple: No juniors today = No seniors tomorrow = No one to fix what AI breaks.
The Path Forward (If We Want One)
The solution isn't complex. It's just uncomfortable.
Accept that quality matters more than velocity. Ship slower, ship working. The cost of fixing production disasters dwarfs the cost of proper development.
Measure actual resource usage, not features shipped. If your app uses 10x more resources than last year for the same functionality, that's regression, not progress.
Make efficiency a promotion criterion. Reward engineers who reduce resource usage. Penalize those who increase it without a corresponding value.
Stop hiding behind abstractions. Every layer between your code and hardware can result in a potential 20-30% performance loss. Choose carefully.
Teach fundamental engineering principles again. Array bounds checking. Memory management. Algorithm complexity. These aren't outdated concepts—they're engineering fundamentals.
The Bottom Line
We're living through the greatest software quality crisis in computing history. A Calculator leaks 32GB of RAM. AI assistants delete production databases. Companies spend $364 billion to avoid fixing fundamental problems.
This isn't sustainable. Physics doesn't negotiate. Energy is finite. Hardware has limits.
The companies that survive won't be those who can outspend the crisis.
There'll be those who remember how to engineer.
What's your organization's response to the quality crisis? Are you optimizing code or buying hardware?
If this resonates, forward it to engineering leaders who need to hear it. Sometimes the most expensive solution is avoiding the real problem.
I spent my entire career in software quality because that’s where thye rubber met the road. My most gratifying stretch was when I had a manager who told developers, if she says it doesn’t go out, then it doesn’t go out—you fix it. Yes, we started working on some tools to automate some of the QA checkin,. But I maintained that software would not solve the “bugs that are “hard to find,” the corner cases. That would take a creative human brain. I eventually quit over such issues. Completely quit—the company (a very large, very well known one), the discipline. I became a computer science teacher, and taught my “value beliefs” along with the details of coding and getting something to work right for a user. I have railed about testing problems for years, still do.
The pipeline problem isn't just a problem in one niche industry. Its a problem everywhere, and the timeline is actually quite a lot shorter than people think.
There are many problems with AI, you've discussed a number of them quite well. What you don't really touch on is the timeline which comes down to economics. If you are not economically compensated, and there is no perception that you ever will be compensated for something, no one goes into the industry. No one invests at the individual level.
AI is a ponzi that eliminates value. It does so in the same ways that Mises write's about with regard to the inevitable failures of socialism (defined as a centrally organized economy where the means of production are owned or controlled by the state). Money-printed environments meet the definitional requirements for this and I've seen nothing that refutes the core failure domains he provides. While the route is circuitous and indirect, the incentives of negative feedback systems being turned into positive feedback systems drive this forward and the trajectory is set with the outcome being just a matter of time.
If value is only ever based on potential human action, and only one party (producers) are enriched in an economy (sieving wealth into few hands) while demand for work is reduced to zero. No one gets hired. No one can pay for anything. Distortions warp the very perception and economic calculation fails to chaos.
Demand isn't need. Its the intersection of supply and demand where the perceived market supply crosses at a price point where money will change hands. Adam Smith covers this requirement in an economic context in his books (1776). You can have great need, and no demand if no one can or is willing to pay for something.
AI ensures there is no demand in factor markets both in terms of cost per hour in throughput, and in the transparent imposition of additional costs caused by jamming communications networks with slop to tortuously interfere with legitimate workers and producers finding a equitable match up. The same thing happens in cellular networks with RNA interference.
A good portion of the job market cycles out every 10 years in every industry. People retire, are injured, have life changes, or die. Its a pipeline that constantly empties.
If you can no longer find work because the competition is high and interference suppresses wages, to the point where you have submitted well over 10,000 applications of customized resume, have a decade of direct experience, and it gets to the 2 year mark where you only got 12 interviews (no offers) as a return on that investment of time and effort. The trend is set. (This has happened to multiple people I know since 2022).
What happens? The competent most intelligent people retrain to something with economic benefit. They abandon their experience as a write-off because they have to. Its a sticky psychological decision, and while they won't stop, or turn down a job its still a lost investment. This has already happened in IT starting in late 2022, the 2 year mark was last year; people are retraining.
Resilient Systems break on a lag, Money-printing has decoupled the need to act, this is a cascade failure that's been happening, and I've been talking about it but all communications are jammed with AI sentiment slop because tech systems have isolated everyone from the reality.
At year 5, you have no replacements. You can't really find qualified help. Brain drain is in full swing. Year 7-10, knowledge is lost. Your semi-competent, bottom run people burn out, it all gravitates to the lowest common denominator just like any other industry that is fueled by money-printing (i.e. Education/Government). No replacements. If you take a lens of economics from Mises, you'll recognize this is the ECP in factor markets. The other problems too all stem from money-printing. Cooperation based on ties to money prevent true economic calculation, GDP growth without datacenters stalls (0.1% for the year). Stagflation takes hold, and the only solution is helicopter money; but that's no real solution. The sieves also happen on the producer side, outsized concentration leads to corruption and large impacts from bad decisionmaking.
Money-printing leads directly to hyper-inflation and then to deflation, or straight to deflation when it doesn't exponentially increase. Prices distort. Banking drives all this. If you look at the changes they made to banking in 2020, interest rates no longer control the bad actors in banking. Fractional reserve was silently discarded in 2020 for Basel3 which is a flawed system based in objective valuation or value (something that's been disproved). I.e. its called a capital reserve system, but there is no reserve. They set deposit reserves to 0% in 2020.
The History of Collapse and Money-printing all cover what we are seeing now. The effects are a problem domain where it lag ahead of any indication of the cause similar to avalanches or dams.
Stage 3 ponzi is here, and these are just the symptoms. What happens to food production when you can't exchange anything for food because currency is worthless. Those that print money steal slave labor from those that hold it. People stop having children when they cannot support the raising of them. Lots of serious serious consequences. By the time the average person figures it out it will be too late to act to change outcomes (hysteresis). Intelligent people should have caught this, but there's been a war against the intelligent for quite some time to the point very few exist compared to previous years. They used to raise everyone up around them, but because of sophisticated techniques to fractionate, and disunify the masses at scale and other things (originating in torture which destroys such people's minds through trauma).
The canaries have largely died silently with little notice. Its a spectrum sure, but the concentration has decreased precipitously. There are quite dark times ahead.
I've been thinking this series of events through for quite some time now and its probably the highest probability we've had for malthusian collapse we've ever had in documented history; though time will tell (and that's just my opinion).