When the Building Stops Needing a Crew

Just in the first quarter of 2026, there were 40% more tech-sector job cuts compared to the same window in 2025 (Challenger, Gray & Christmas, 2026). We keep reading about people losing their job or looking for one every single day on social platforms. A friend gets laid off, then another, then a colleague you used to sit two desks away from. These days one would think LinkedIn's gone ecofriendly with all its green "open to work" profiles.
“Ye Ho kya raha hai, bhai” #YehJoHaiZindagi
If you are in the tech industry and one of the short-stick holders, it is but natural for you to be worried.For expats like myself in UAE, the mortgage is still real, the visa end date is still ticking, the kids' school fees are not gonna vanish.
None of that is something you could or should ignore. But, this article is not to empathise, motivate, give hope, or to argue you out of the fear.
I am writing this because I see an alternative storyline that is overlooked. There is something happening underneath the layoff numbers, and the people who see it this year are going to be writing the next decade's stories.
Curious to know what my take on this is? Then keep reading.

The old economics of building

Until recently, building a real digital product needed an army and then some. A frontend developer. A backend developer. A database person. A designer. A DevOps engineer. A QA tester. A product manager to keep the whole thing pointed in roughly the same direction.
The most agile team was six people minimum, and that was just the development team for something small. Most companies I have worked with would call that a lean team.
The reason was simple. Each of those roles required years of specialised training. You could not credibly do all six. The economics of building forced you to assemble a team or work for one. If you were one of these specialists and wanted to build something, the first step was finding five other people.
That was the deal. But not any more.

Rise of the Specialisation Assistant

A single builder today, with the right toolkit, has an AI assistant (Agent) for every one of those specialisations. The frontend assistant. The backend assistant. The database assistant. The design assistant. None of them is as good as a senior specialist with fifteen years of experience. None of them needs to be. They are good enough to get a working product into a customer's hands, and the cost of running them is a rounding error against what one senior engineer's salary used to be.
And don't even get me started on the turnaround time.
I call this the rise of the Specialisation Assistant, but I am sure someone has already said it before. We digress.
The point I am trying to make is that the specialisations have not gone away. The need for skill, taste, and judgement at the top of the stack has not gone away. What has gone away is the requirement that the specialisations live inside separate human bodies, each charging a salary, each needing to be hired, managed, and retained.
When you read "you don't need developers anymore," that is the wrong reading of the trend. The right reading is "I don't need an army of people to build my project anymore so I am not hiring or firing." Those are different sentences, and only one of them is being printed.

The SaaS bet, and why it is being re-examined

For most of the last fifteen years, the right answer in enterprise tech was clear. Don't build, buy.
SaaS has slowly become the defacto for every category of enterprise product thanks to the rise of enterprise cloud infrastructure. The math worked. A SaaS vendor amortised the build cost across thousands of customers, and any one company trying to build its own version was paying multiples for the same capability. Owned applications were treated as legacy thinking. The CIO who said "we build it ourselves" was the CIO who was about to get replaced.
I have lived inside that consensus for years. I have argued for it. I have helped clients and managements walk away from custom builds and onto SaaS rails. The math was solid, and I have years of successfully delivered projects and products to vouch for it.
But with every passing day I feel that AI driven "Specialisation Assistants" are what will finally break the math.
Today, one builder plus a toolkit can produce what used to be the work of a whole team. With the cost of infrastructure dropping, the cost of an owned, tailored application if done right can be competitive to the cost of a SaaS subscription.
The custom-designed CRM built around the actual shape of your business is no longer ten times the price of the off-the-shelf one. It might be the same price, or slightly more expensive, but it offers you control and ownership not only of your data but of the entire ecosystem, including what features your stakeholders actually need instead of what the vendor decided to ship. It gives you control back, and that is priceless.
Don't just take my words for it.
A 2026 survey of 817 enterprise builders by Retool found that 35% of teams have already replaced at least one SaaS tool with a custom build, and 78% expect to build more custom internal tools in 2026 (Retool, 2026).
The wind is moving back toward owned applications, custom-fit, built around the work rather than around the vendor's roadmap.
This is not a nostalgia move. The reason the industry chose SaaS was never that SaaS was inherently better. It was that the build economics did not work. Now they do. So the choice is being re-examined.
This is the practical answer to "what should I be learning right now." Not how to ship another SaaS tool into a saturated market. The practice worth getting good at is building tailored solutions for organisations that are quietly walking back their SaaS sprawl. That is where the work is going.

A Tale of Two Cities (startups)

Let me tell you a story of two companies with similar DNA and very different outcomes to sink this in.
Builder.ai and Base44 are both app-building companies that pitch the same thing. Describe what you want in plain language, get a working application.
Builder.ai launched in London in 2016 as Engineer.ai, promising to make app development as easy as ordering a pizza. It raised roughly $445 million from Microsoft, the Qatar Investment Authority, SoftBank's DeepCore, and others, and reached a peak valuation of about $1.5 billion (Olson, 2025). It employed around 1,500 people. It claimed that the apps were built using its AI engine, Natasha, but the trouble was that the underlying AI never quite did what the marketing promised. A 2019 Wall Street Journal investigation found that the company's AI capabilities were significantly exaggerated, and that much of the actual coding was being done by human engineers in India and elsewhere (Purnell & Olson, 2019). In May 2025, after a creditor seized $37 million from its accounts, the company entered insolvency proceedings, owing tens of millions to Microsoft and Amazon (Bloomberg News, 2025).
In February 2025, while Builder.ai was already inside the spiral that would end it, a 31-year-old developer named Maor Shlomo opened his laptop in Tel Aviv and started building Base44 alone. The product pitch was almost identical to Builder.ai's. Describe what you want, get a working application. Shlomo bootstrapped it with his own savings. He hit one million dollars in annual recurring revenue three weeks after launch. He grew the product to 250,000 users in six months, and by the time of the acquisition he had assembled an eight-person team. In June 2025, four months after Builder.ai filed for bankruptcy, Wix acquired Base44 for $80 million in cash (Bort, 2025).

The company that pretended to have AI raised $445 million and collapsed. The company that actually had it was bought for $80 million in cash a month later.

The technology that Builder.ai marketed as AI but had to deliver through human engineers nine years ago, Base44 actually had in 2025. That is the gap that closed. That is what the Specialisation Assistant looks like when it stops being a marketing claim and becomes the way the work actually gets done.

In a single month in 2025, Builder.ai died with thousands of developers losing their jobs, while Base44, built by a single developer who scaled up to a small team and their "specialist assistants," was bought for cash. The difference between the two companies is what I think is the real story.

What it asks of you

I am not selling you a clean future.
The toolkit is here, but the toolkit alone does not build anything. Most of what gets built will fail. The moat ahead is not the tech stack, which is now available to everyone. The moat is in the real understanding of the customer you are building for and the value you can provide to them. Those are harder to acquire than a credential, and not everyone who was a good coding specialist will turn out to be good at them.
If you are reading this and thinking "that does not sound like me," that is useful information. The future landscape rewards a specific temperament. It is not better than the old one. It is different.
But if you have a problem you actually care about solving, and the discipline to sit with it long enough to ship something useful before you get bored, this stretch of years is the door you have been waiting for. Whether you knew it or not.
The headlines this year will be about the layoffs. The headlines five years from now will be about the people who used this stretch to build the thing they had been carrying around in their head.
Whether you are in that second group is, for once, almost entirely up to you.

References

Bloomberg News. (2025, May 20). Microsoft-backed Builder.ai set for bankruptcy after cash seized. Bloomberg. https://www.bloomberg.com/news/articles/2025-05-20/microsoft-backed-builder-ai-to-enter-insolvency-proceedings

Bort, J. (2025, June 18). 6-month-old, solo-owned vibe coder Base44 sells to Wix for $80M cash. TechCrunch. https://techcrunch.com/2025/06/18/6-month-old-solo-owned-vibe-coder-base44-sells-to-wix-for-80m-cash/

Challenger, Gray & Christmas. (2026, April 2). Challenger report: March cuts rise 25% from February, AI leads reasons. Challenger, Gray & Christmas. https://www.challengergray.com/blog/challenger-report-march-cuts-rise-25-from-february-ai-leads-reasons/

Olson, P. (2025, May 23). Microsoft-backed Builder.ai bankruptcy is a problem for AI washers. Bloomberg Opinion. https://www.bloomberg.com/opinion/articles/2025-05-23/microsoft-backed-builder-ai-bankruptcy-is-a-problem-for-ai-washers

Purnell, N., & Olson, P. (2019, August 14). AI startup boom raises questions of exaggerated tech savvy. The Wall Street Journal. https://www.wsj.com/articles/ai-startup-boom-raises-questions-of-exaggerated-tech-savvy-11565775004

Retool. (2026, February 17). The build vs. buy shift: How vibe coding and shadow IT have reshaped enterprise software. Retool. https://retool.com/blog/ai-build-vs-buy-report-2026