Why NOT to Panic About Tech Layoffs

10 Positives About the Future of the Software Industry

As we enter the last month of the year I’m actually quite optimistic about the potential growth of the software industry in 2026 and beyond.

I’ve spent the last 20 years as a CTO watching wave after wave of innovation reshape the industry, and while AI is genuinely different than other technologies in a few important ways, the broader pattern is familiar: disruption makes people panic at first, then things settle down & start growing again.

Here’s an example:

It was 2008 & I had a close Sys Admin friend who worked at a famous U.S. company that was doing all their hosting at Equinox’s data center.

This person was really panicking about this “new thing” called AWS. To my friend, Cloud Computing was mysterious and probably meant the end of his career.

He imagined all kinds of horrible scenarios where AWS & its fancy “web console” would take over his job maintaining a data center full of servers.

And to some degree this did happen. The nature of sys admin work did actually change because of AWS & Cloud.

But over the years, due to Clouds’ new way of doing things, there was also much more demand for new skillsets which kept sys admin job growth rapidly increasing.

On top of that many of the skills my friend thought would disappear (like knowing how to use commands like “grep” or having a good understanding of Linux services) are still around today.

So in the end like 10s of thousands of other sys admins he was fine and still works to this day as a VP CloudOps somewhere.

The point is that with all of its amazing powers, AI will still generally create more jobs in the long run than destroy them, but right now everyone from shareholders to CEOs to individual developers are panicking.

Fear is always the 1st reaction to new innovation.

So let’s look at some reasons NOT to panic:

Don’t Panic: Software Is Still Eating the World

The birthplace of AI is Silicon Valley. If you add up all the software companies there, they produce about $2 Trillion in revenue. Meanwhile, the GPD of the U.S. alone is something like $30 Trillion.

Meaning there are massive industries and segments that software has barely touched including transportation, manufacturing and physical goods.

These industries are very much behind the times when it comes to software & tech, so there is plenty of greenfield to pursue in the next decade.

Don’t Panic: Industries Have Huge Needs

Industries like HealthTech, FinTech & EdTech are massive & complex. Their need for software solutions is therefore also significant & getting bigger every year.

For example, if you count up just the number of B2B software companies in the Health Care / Life Sciences space alone, its easily in the range of 10,000 in the U.S. That number has been growing year over year for 20 years & will continue to grow with demand.

In fact, AI will probably just explode that number as companies start to see the potential of AI to help their customers.

Don’t Panic: AI Still Underperforms

Show me a clear case study of an average company using AI to significantly improve a business metrics like revenue, profit, margin, NPS, etc. It’s extremely rare.

AI has unbelievable potential, but right now it still fails at many basic things. It is often totally innaccurate. It struggles with multi-step reasoning. It falls apart with bad prompts or incomplete context. And it requires a massive amount of scaffolding to make it useful in a real enterprise workflow.

That means AI can’t replace most engineering jobs. A powerful tool, sure, but still one that needs humans. Ask AI to look at a large, complex, 10 year old codebase and then assist you with writing code or tests. The result isn’t great.

Don’t Panic: AI Adoption is Slow

I talk to a lot of CEOs and one of them said to me, “Bobby, I don’t care if they get to AGI this year, we just don’t have time for it.”

Think about what he said: no time for even AGI. The issue for companies is not technology, its time.

Companies have too much other stuff going on like dealing with competition, generating sales, marketing, etc to adopt AI quickly. What’s happening right now is non-systemic adoption of AI. Meaning, in a random company you’ll have everyone using AI for personal use (i.e. shadow AI). But they typically won’t use it systemically in major business-outcome driving projects.

That means there is plenty of room & time to help companies adopt AI & learn how to leverage it.

Don’t Panic: AI Companies Exaggerate

This is the odd part about the AI revolution. AI companies like OpenAI and Anthropic greatly exaggerate their capabilities. I don’t remember AWS or even social media companies doing this. Their tech was what it was.

Sure, every company wants to claim their tech is great, but what we hear from the big AI companies is pretty outlandish.

Remember when GPT5 was supposed to be revolutionary? Well, in many ways it went backwards… I don’t doubt we’ll have another transformers-level breakthrough one day, but that could easily be 10 years away.

In the meantime, we have to live with the AI of today, which is not that different from what it was at the start of this year.

Don’t Panic: Regulations Will Come

Every major technological shift eventually hits a regulatory wall. And AI will be no different.

Governments move slowly, but they do move, and they always show up late but strong. Once the regulatory frameworks arrive, companies will be forced into structured, safer, slower adoption cycles.

That means guardrails, certifications, compliance steps, required human oversight, and entire new roles dedicated to managing AI responsibly.

Far from accelerating job loss, regulation creates whole new categories of work like AI governance, AI auditing, AI risk, AI compliance, etc and extends the timeline for actual disruption.

Don’t Panic: Industry Culling Was Needed

Let’s be honest, the tech industry was super bloated. Massive teams. Too many PMs shipping nothing. Too many young kids graduating from college with CS degress who were never meant to be programmers. And, too many layers of management. In a typical $100M software company today you can easily have 50 to 75 engineering management staff.

The layoffs we’ve seen over the last two years is reseting the bar. Companies downsized, reorganized, and came out leaner.

And historically, the industry always rebounds harder after these corrections. Dot-com bust. 2008 crash. Post-pandemic. Same pattern: cut, then stabilize, then grow faster than before.

Don’t Panic: AI Encourages Creativity

Ironically, the biggest thing AI accelerates isn’t job loss it’s experimentation. Developers & non-developers try more ideas because AI lowers the cost of prototyping.

Teams attempt features they would’ve never greenlit before. Entire categories of tools, plugins, and integrations get built simply because the overhead to test them is getting closer to zero.

Creativity always expands when friction to build drops. And expansion means more projects, more roles, and more “surface area” across the entire software ecosystem.

Don’t Panic: Jevons Paradox

Here’s the part most people don’t understand: when something becomes more efficient, we don’t use less of it — we end up using more.

That’s Jevons Paradox. AI makes developers faster, but that doesn’t shrink the total demand for software. It increases it. Every time efficiency goes up, consumption goes up even faster.

The world doesn’t want fewer apps, systems, workflows, automations, or digital experiences. It wants exponentially more and AI just pours gasoline on that demand.

Don’t Panic: AI is Still a Baby

Despite the hype, AI is in its infancy. It can generate code, summarize text, and answer questions … but it still can’t reason deeply, manage context over long horizons, or operate reliably without human supervision.

It breaks easily, hallucinates, and needs enormous amounts of tooling and glue code to behave. In engineering terms, this isn’t a mature platform, it’s a prototype with incredible promise. And prototypes don’t replace entire workforces.

They evolve, slowly, through years of iteration, standards, best practices, and infrastructure improvements. We’re at mile one of a hundred-mile race, which means there’s plenty of time for people, companies, and entire industries to adapt.

Conclusion

We’re in a transition period in the software industry & transitions are difficult, there’s no doubt about that — we would all rather be in the “easy, stable period.”

But the reality is that we’re only 2 years into the AI-era and because its so shrouded in hype it doesn’t make sense to make too many predictions of doom.

In fact, if you listen to some of today’s AI evangelists, you’d think we’re six months away from sentient software running every company on Earth.

But we’ve seen this movie before.

In the 1950s, futurists confidently predicted flying cars, robotic maids, moon colonies, and cities in the sky by the year 2000.

None of it happened.

What actually materialized were incremental, practical innovations that took decades to mature. Now, AI won’t take decades but 2 years is certainly not enough.

The loudest predictions are almost always wrong because they assume linear, uninterrupted progress & ignore the messy middle of enterprise organizational structures, infrastructure changes, human behavior, change management & regulation.

The future never arrives as fast or as cleanly as the hype machine claims!

Hope you have a great end of year and keep the shark swimming! :) 🦈