AI Made Me Homeless

Economy

Software Engineering

Technology

AI Made Me Homeless

How a $500k/Year AI Engineer Lost Everything

Eric Elliott

12 min read·

2 days ago

--

Three years ago, I was earning $500,000 a year in tech. Today, I’m homeless. I live in my car, where I keep most of my belongings and try to get enough sleep to function. I have a cheap gym membership so I can shower. I don’t have any meaningful income, and I don’t know when I’ll find work again.

I’ve written three books on software engineering. I’ve helped build products used by millions of people, including Adobe Creative Cloud, Adobe Firefly, Google Search, and YouTube. I’ve led engineering teams, built successful businesses, and helped startups hire and train their engineers.

I’m a Webby-nominated AI engineer who built my first neural network in the 1990s. I’ve spent decades learning, building, and adapting to new technologies. When people tell displaced workers to “just learn AI,” they’re missing the point.

I already know AI. I helped build the AI revolution that’s displacing us.

“Just learn AI” is not job security. I want to tell you how I got here, because I think what’s happening to me is about to happen to millions of other people. This isn’t just a personal story about falling on hard times. It’s a warning about the economic consequences of AI.

For years, I ran a successful software consulting and developer mentorship business. My days were filled with mentoring developers, teaching them how to code, how to build great software, and how to run great software engineering organizations. We had individual developers who invested in weekly mentorship to advance their careers. We also had well-funded startups that hired us to train entire engineering teams, recruit engineering leaders, and help build their engineering organizations. One of our clients was Hopin, one of the fastest-growing startups during the pandemic. We helped recruit a key engineering leader for the company.

Some of the developers I mentored eventually became colleagues. I’d help them develop their skills, they’d graduate from the mentorship program, and sometimes I’d hire them to work on my teams. One developer I mentored ended up working with me at three different companies. I loved that work. We were helping people build careers and helping companies build better software. It was rewarding, meaningful work, and the business was doing well. A big part of our success came from my technical writing.

At its peak, my blog was getting approximately 700,000 reads per month. Those articles were our primary customer acquisition channel. Developers would discover my writing, learn from it, and some would eventually become mentorship or consulting clients. It was a business model that had worked for years.

I’d been using AI long before ChatGPT became popular. I understood what the technology could do, and I was excited about it. But before ChatGPT, relatively few developers were using AI tools regularly. They weren’t having a substantial impact on our business. Then ChatGPT launched, and suddenly millions of people had access to an AI tutor and coding assistant. Why pay someone to teach you how to code when you can ask an AI? Why hire a mentor when you can get answers to your questions instantly?

Human mentorship still has tremendous value, but the economics had changed. And something else happened that was even more devastating. Our blog traffic collapsed. We went from approximately 700,000 monthly reads to 70,000 within a few months. Then 7,000 a few months after that. We lost approximately 99% of our readership in a matter of months. Our primary customer acquisition channel had effectively disappeared. We still had happy customers. Some developers stayed in our mentorship programs for years. But eventually, people graduated. They learned what they came to learn, advanced their careers, and moved on.

That was always part of the business. The problem was that new customers weren’t coming in to replace them. I tried adapting. I experimented with different content formats, including video. I’d had successful videos before, and I thought maybe I could reproduce that success and build a new customer acquisition channel. But nothing worked reliably. I kept trying new things, hoping something would catch. Meanwhile, revenue kept falling. Eventually, I realized we weren’t bringing in enough new customers to sustain the business.

I started looking for engineering leadership positions. At first, I was just feeling out the market. I’d spent decades building software, leading teams, writing books, and teaching engineers. I thought finding a management position would be relatively straightforward. It wasn’t. As the search dragged on, I realized I was facing a much more difficult market than I’d anticipated. I went all in on the job search. In total, the process took roughly a year. During that time, the business continued to decline, and I was drawing more and more from my savings to keep us afloat.

I was scared. What we’d always done wasn’t working anymore. I didn’t know what to do next, and I didn’t know whether I could adapt quickly enough. Apparently, I couldn’t. Eventually, I landed an engineering management position at Adobe, working on Adobe Firefly. I was relieved. I’d worked at Adobe before, and I liked the company. I thought maybe this was my opportunity to rebuild some financial security. Maybe I could retire at Adobe.

Taking the job meant relocating from Los Angeles to the San Francisco Bay Area. Finding housing was difficult and expensive. I spent a lot of time in hotels while trying to get settled, and the moving expenses consumed much of what remained of my savings. By the time I started at Adobe, my bank account was nearly empty. But I had a good job again. I could rebuild. At Adobe, my team and I took the initiative to establish Firefly Client Platform: a shared component system that made it easy to add AI features to any application surface.

It was meaningful work on one of Adobe’s most important AI products. My manager had me on track for a promotion, but he left the company. The new leadership had other plans before they’d even met me.

I’d spent two years trying to rebuild the financial security I’d lost when my business collapsed. Now I was unemployed again, with almost no savings, facing the same brutal job market.

I returned to consulting. The original mentorship business model was no longer working, so I started experimenting with new services and offers. I worked tirelessly to find clients. Then our lease ended, and our landlord wanted to move into the house. We had to find a new home, and the move was expensive at a time when we could barely afford it. Shortly afterward, I landed a substantial consulting engagement. For a moment, I thought I’d found my way out.

If I could land one client, I could land more. We worked hard to take good care of that client. I expected the engagement to continue through the third quarter. Instead, it ended near the beginning of the second quarter. We lost approximately 90% of our revenue overnight. We’d had about three months of good income before it disappeared again. In the meantime, I hadn’t been able to recruit enough additional clients to replace it.

I picked up smaller consulting projects here and there. Building prototypes, implementing features, helping startups solve technical problems. But the work wasn’t consistent. It was enough to occasionally put food on the table. It wasn’t enough to pay the rent. Without enough income to cover the rent on our new home, we fell behind.

Being evicted with nowhere to go was terrifying. I grew up in a modest household. My dad was a great guy, but sometimes he worked two or three jobs just to keep a roof over our heads and food on the table. Even then, we sometimes needed help. My family doesn’t have money. I have friends and professional acquaintances, but I don’t have a network of wealthy people who can afford to support me indefinitely.

There was no financial safety net waiting to catch me. We spent some time in cheap motels. Then the motel money ran out.

It’s where I sleep. It’s where I keep my belongings. It’s where I try to get enough rest to continue searching for work. I have a gym membership primarily so I can take a shower. A friend has helped us by letting us keep some belongings at his place. We also have a storage unit several hours away. Everything is scattered. Every day is about figuring out how to survive while trying to find a way back into the workforce.

I’m still applying for engineering leadership positions. I’ve had promising interviews. I’ve made it through final rounds. But the offers aren’t coming. And I don’t know when they will. That’s a frightening place to be when you don’t have money for a hotel room, let alone an apartment.

Here’s what I really want people to understand. This isn’t just a story about one successful engineer falling on hard times. It’s a systemic problem. I happen to be on the front lines because I work in software engineering, one of the first industries to experience widespread disruption from generative AI. But the same forces are coming for other industries. Companies are learning to do more with fewer people. They’re investing heavily in AI while reducing headcount.

And the labor market is becoming flooded with extraordinarily talented people competing for a shrinking pool of opportunities. These aren’t necessarily underperforming employees. They’re experienced engineers, architects, managers, and technical leaders who have built some of the most successful products in the world. They’re losing their jobs because companies believe they can operate with leaner teams. Consider the scale. In 2026 alone, US technology companies have announced roughly 170,000 job cuts. Microsoft recently offered voluntary buyouts to approximately 60,000 employees.

Those buyout offers aren’t confirmed layoffs, but think about what happens when your employer offers you a buyout. You take it and start looking for another job, as do many of your colleagues, because it’s a very clear signal: your job is at high risk. Suddenly, tens of thousands of additional experienced professionals are competing for the same shrinking pool of jobs. And the competition for remaining positions is intense. People are submitting hundreds of applications.

Experienced engineers are struggling to get interviews. Even people who make it through multiple rounds of interviews aren’t necessarily getting offers. I’ve spent nearly three decades working professionally in software engineering. I’ve written three technical books. I’ve contributed to products used by millions of people. I’ve managed engineering teams, built businesses, mentored developers, and helped companies ship successful software.

If someone with my experience is struggling this much, what happens to everyone else?

One common response to job displacement is entrepreneurship. If companies aren’t hiring, build something yourself. And AI has made building software dramatically easier. People who have never written code before can now create working applications using AI coding tools. That’s extraordinary. It’s also creating another economic problem. The barriers to building software are collapsing. But the barriers to finding customers haven’t disappeared. App stores were already brutally competitive before generative AI. Most apps never become commercially successful.

Now we’re entering a world where millions of people can build applications with minimal technical experience. And those people are legitimate competition. I can build software. I’ve been doing it professionally for nearly three decades. But what should I build? What product will people pay for? How do I reach customers in a market flooded with new applications? And what happens when potential customers can simply ask an AI to build something for them? These are difficult questions even for experienced entrepr