Software engineering is in rapid decline. Or is it?

That's certainly the story being told. We've all seen the headlines - mass workforce reductions, announcements from the likes of Oracle, CNN, Dropbox, Block and Amazon slashing engineering staff citing optimisations through expanded use of generative AI and agents.

Just in March '26 alone, two weeks into the month, we'd already had announcements from Atlassian and Meta, cutting thousands of roles as they look to leverage AI to cut costs.

But is that reality or a veiled excuse? Perhaps it's something in between? What does this all mean for the future as engineers? If I were heading to university now, would it still be worth pursuing a career in software engineering, would I be able to find a job?

I'm going to look at this from two angles as I think there's a distinction to draw here between current positions and future ones.

• The current positions and mass layoffs - the jobs that are filled today, why engineering teams are facing massive downsizing, is AI the real underlying reason or could it possibly be something else?

• Future hiring and the recruitment market - is the number of advertised engineering jobs in decline? Is the profile changing? Is this still a career worth pursing?

Layoff announcements - AI Washing for corporate downsizing

If we were to believe all these AI related layoff announcements, you'd expect me to say something scary like "AI will cause 75% of all engineers to be out of work by 2027" - but I'm not going to be saying that. (yet!). I have a hard time stacking the numbers up for these announcements.

Is AI replacing you today?

The scale of layoffs doesn't match the current capabilities of AI. We're seeing engineers laid off en-masse often in the thousands and accounting for as much as 50% of the team being laid off in single rounds, yet current estimates suggest AI agents can only fully replace a very small fraction of an engineer's entire role - low single digits in controlled studies. For example, recent studies (e.g. Centre for AI safety / Scale AI - https://scale.com/blog/rli) suggests only 2.5%.

That's not to say AI augmentation doesn't massively increase productivity, it absolutely does - the parts it's good at, it'll boost significantly, but it's a massive boost to a smaller portion of the role as a whole. It's nowhere near capable of replacing an engineer one-for-one, no matter what headlines you read.

Because the role isn't just about writing code, which is yes, largely a solved problem when one or more AI agents is working as a team with an experienced engineer, but being an effective engineer spans the 5 dimensions of the "SPACE" framework, not just how many lines of code you write.

The SPACE framework is an attempt to answer the question "What does it actually mean for a developer to be productive" - it was introduced by researchers from Microsoft, Github and academia to move the industry away from crude metrics like lines of code or tickets closed. It reframes developer productivity from how much code is written to how effectively teams deliver value - sustainably, collaboratively and with minimal friction. https://www.microsoft.com/en-us/research/publication/the-space-of-developer-productivity-theres-more-to-it-than-you-think/

When considered in the context of SPACE, measuring the actual impact of introducing AI can't be reduced to a simple "x% productivity gain" - productivity is multi-dimensional and context dependent. AI primarily improves activity and efficiency - the easiest parts of the job to measure - but has far less clear impact on performance and collaboration, which are the dimensions that actually determine outcomes.

There's gains to be had for sure, but they aren't at the "layoff 50% of engineering staff" levels - something doesn't stack up.

So why so many layoffs?

Spoiler - AI has some impact, but it's mainly traditional corporate downsizing due to over-hiring.

Whilst borrowing was cheap we saw companies increase headcount at pace. Not because they needed 1000 engineers today - they hired because they expected to need them in three years - planning for growth far beyond the immediate horizon.

They over-hired whilst adding layer upon layer of hierarchy, they duplicated teams, created inefficient silos, until they're bloated to the point that productivity slows and cohesion fails.

AI or not, when market conditions change unfavourably, that bloat needs to be cut - and that's traditional corporate downsizing.

The January '26 challenger report stated that, across all job markets in the US, 108,435 jobs were slashed in total, and 7624 of these cited AI as the reason. (https://www.challengergray.com/blog/challenger-report-january-job-cuts-surge-lowest-january-hiring-on-record/).

Even in a record month for layoffs in Jan '26, only 7% of cuts were explicitly attributed to AI - the majority were driven by economic conditions, restructuring and contract losses. That's not what you'd expect if AI were the primary driver of layoffs.

So why cite AI?

Spoiler again - market messaging and stock prices.

Corporate downsizing affects confidence - it paints a picture of a struggling company, or that bad decisions were made leading to bloat, which would normally have some negative effect on stock prices. But, dress it up in AI efficiency clothes and it's no longer a course correction, it's a positive to that companies overall bottom line and the stock price goes up.

While some studies show modest declines in employment and job postings since the release of ChatGPT, multiple studies, including Goldman Sachs and NBER, indicate that highly AI-exposed occupations have not experienced significantly greater contraction than the wider labour market.

The bottom line is there's real evidence of reduced demand in AI-exposed jobs (which I'll come back to in a moment), but also strong evidence that AI-exposed sectors aren't collapsing faster than others.

So, for me, the only conclusion I can draw for these large layoffs is "AI washing" - spinning a positive message for the markets on something that was actually inevitable corporate downsizing. What we're seeing isn't AI replacing engineers at scale - it's companies using AI as a narrative to justify decisions they would have had to make anyway.

The future jobs market and new engineers

Whilst I believe that the current lay offs are nothing more than corporate downsizing wearing an AI coat, there is very real evidence of a shift in demand for engineers and the skills being sought.

There's some bad news (and opportunities!) here for Junior and entry level roles, and some good news for more senior positions. Overall, demand for engineers hasn't disappeared- it has become far more selective - favouring experience and AI adoption.

According to the CompTIA Tech Job Report, active job postings in the US, for tech occupations, increased 9% in February alone to 505,045 active vacancies. The total count is still roughly flat compared to pre-pandemic levels, but is increasing month on month.

But, the profile of those roles is changing.

Engineers are still in demand, but the demand profile has changed.

First off, hiring has become more senior focussed. Indeed's september 2025 labour-market analysis (https://www.hiringlab.org/2025/09/25/september-labor-market-squeeze-on-new-entrants) found that software development had significant senior skew (of any sector) with 30.7% of postings carrying senior titles and only 2% carrying junior.

The other 67.3% isn't evenly distributed across levels, it appears to skew toward mid-level engineers or senior roles that don't explicitly call out senior - e.g. "full stack engineer" and "5-8 years experience".

Very few roles are explicitly targetting entry level positions, while a large proportion explicitly target seniors, the rest are ambiguous - but rarely entry-level.

Alongside this senior skew, we're seeing a rise in the requirement for AI as a key skill to match the job profile. In the US, more than 20% of postings opened as of December '25 mentioned AI, in the UK this was closer to 6% (to the end of October '25) and a similar 6% of postings in Canada (to the end of November '25).

The market for experienced engineers is certainly still there, but employers increasingly want engineers who have a demonstrable ability to operate at a higher level - systems thinking, product judgement, architecture and AI-assisted delivery - not just people who can write code.

That's all good news for established engineers, tempered by the bitter taste that it's also bad news for the juniors coming up behind us - but I think I can soften that blow a little, once I've talked about the weirdness in the UK market, where things aren't quite so rosy.

The UK anomaly

In the UK things are a little more bleak. The UK job market overall is still softer than pre-pandemic levels - and software engineering has been hit significantly harder, with some estimates suggesting tech roles are still more than 40% below 2019 levels. (ft.com)

Whilst the US market is stabilising and showing signs of recovery, the UK market remains materially depressed, particularly in software engineering. Why?

The divergence between the UK and US isn't accidental. The US remains the centre of gravity for AI investment, big tech, and venture-backed startups - all of which continue to drive demand for engineers, even amid layoffs.

The UK, by contrast, has fewer large tech anchors, an arguably smaller startup ecosystem, and is more sensitive to macro-economic conditions. The result is a different dynamic - the US is seeing reallocation of engineering roles, the UK is experiencing a genuine contraction, particularly in advertised software engineering roles.

Whilst it's bad news for juniors, it's also an opportunity

I've covered the stats already, and it's clear that junior positions are becoming few and far between - not because engineering is disappearing, but because the nature of entry-level work is changing.

AI is particularly effective at the kind of routine, low-context tasks that used to form a large part of junior roles, which means companies need fewer traditional entry level positions, especially in a tighter hiring market.

So what can the new entrant do to increase their chances of landing one of these roles, or even leapfrogging the role to a mid-level position? Is it even still worth studying computer science and pursuing software engineering as a career? What advice would I give an 18 year old thinking about this career path or to the newly graduated 22 year old who expected to land their first position this year.

Let's start with the degree. Do it.

The nuanced part is - as a hiring manager, I don't actually care if you have a degree or not. Some of the best engineers I've worked with over the last 30 years are self-taught and don't have one. That demonstrates to me a desire to learn and grow, to get grounded in the fundamentals, to pursue a craft that is part creative and part scientific. A degree on the other hand gets credibility and through the basic screening processes, it also establishes knowledge in those foundations and demonstrates you're ready for the next phase - practical application.

But is it worth even pursuing? Yes, absolutely.

Software engineering is changing rapidly, but the stats show the roles are still out there - they're just different to what they were 12-18 months ago. It's still a high reward career choice with the fun of complex problem solving usually tied to delivering real impact to peoples lives. I see nothing to suggest at the moment that this is going to change any time soon, so it's absolutely worth pursuing, just be ready to evolve professionally at the same pace as the role is evolving.

Getting on the ladder is the hard part - landing that first junior role, or leapfrogging to a more in-demand intermediate position. There's things you can do here - and the answer isn't just using an AI to fire off your CV to 4000 vacancies.

Start now - build something. Start your own projects, contribute to open source, get involved in a project on github or start your own. Demonstrate the practical application of your skills to employers, don't just wait to land a junior position.

Getting started and building your own software, own service, maybe even launching it not only gives you some commercial exposure and experience, it might even set you off on a different track - that of the solopreneur.

That AI stuff that is being banded about so often as reason for layoffs - it's there for you too. You have something previous generations didn't - an entire team of AI agents at your disposal to help you build out your own services, to disrupt existing stagnant markets, to out-pace other companies that aren't keeping up. It might pan out, it might not, but it's valuable experience either way. If it pans out, you'll find yourself a founder, potentially with multiple income streams and if it doesn't, you've created experience that might help you land the job you're looking for, or even leapfrog to a more in-demand intermediate position.

Conclusion

No matter which way we cut this, it's a bleak picture. Layoffs aren't pleasant at all and there's a lot of them happening currently, but we can take comfort that it's not caused by AI replacing engineers just yet. The market is becoming more AI focussed, more selective, but it's shifting, not shrinking.

Rather, AI is an opportunity. It's a crucible for the next generation of start ups and engineers to evolve, to improve, to build more and build it better and faster than ever before and it still needs the seasoned engineer at the helm.

The role of the engineer isn't disappearing - it's being redefined. The job market isn't collapsing, it's becoming more selective, and for those willing to adapt, there's never been a more powerful set of tools to build with - if you know how to use them.