It has been a wild couple of weeks to be anyone who touches software.
First it was PHP2Shell. Then Node2Shell. Then it felt like everything to shell, with nearly every stack vulnerable at the same time. Every morning you woke up and there was more fallout. (If you’ve been under a rock for the last week, SecureLayer7 has a good rundown of the whole “2Shell” family, including how they connect back to Log4Shell.)
We’ve known for a while that a tidal wave of vulnerabilities was coming. Some people have been fighting that fight for a long time already, but this felt like the first really big one. It won’t be the last.
Right now it’s a hard time to work in security. Everyone has a flak jacket on, taking the punishment and working the front lines. It’s tough.
I remember the morning Log4Shell hit. I was at a healthcare company with a very small appsec group. A couple of my people were on PTO, and watching the app firewall wasn’t even officially part of my job.
I woke up, was reading tech news, and saw what was happening. It dawned on me: wait, this is my problem. I had to figure out how to protect my company from this specific issue, and I had to jump into action fast.
It was a tough day. It was also an exhilarating one. Knowing you may be the single person who stops, or doesn’t stop, an attack that big is unnerving.
We’re getting a lot of signals from developers too about what life is like right now.
James Cross runs a YouTube channel about becoming a software developer. He’d gone quiet for a while, and this week he posted a video explaining why. He hasn’t felt right giving tips on breaking into software when the industry may not exist in a couple of years, and if it does, may not be a good one to be in.
In the video he tells how he went from construction worker to software developer, then spent extra time getting better and better to climb the ladder. That is the software developer’s story. I know so many people who made that move from another career. They worked all day, then went home at night to read articles, watch videos, and build side projects, grinding to get better at the craft.
And there was a great path waiting for them. You could go super deep in technology, or move into people management or project management. For a long time, it was a really great career.
Now we’re watching middle managers get removed and developers get laid off, while AI changes the entire scope of the job.
For the people left after the layoffs, the expectation is that AI is here, so there should be less work. That isn’t what’s happening. Most of the developers I talk to are seeing their workload go up.
Companies cut developers on the theory that each one can now do more, and we’re back in the Mythical Man-Month. One plus one does not equal two, and that goes for agents as well. Somebody still has to point the agents at the work and check what comes back, and that’s the developer who’s still there.
So there’s a lot of discouragement out there right now. If you touch software right now, it is not a fun time to be doing what we’re doing.
And the vulnerabilities are only part of it. This happens whenever we have revolutionary change. The tidal wave doesn’t ask the beach for permission to hit it. All we can do is get to high ground and clean up what’s left over. I don’t know where we are in that wave. Is the water still going inland? Is it just about to hit? It’s hard to know what the next few years look like.
But people are looking at their careers and asking: why am I trying so hard when it won’t matter? Why should I keep fighting for a company that will lay me off the moment AI can do my job?
Those are the questions that stop a profession. Work becomes monotony, a job you clock in for and clock out of. We’ve always had some of that in software, but what I loved about it was that it was a place for creators. People were passionate about what they built. It was a puzzle, and I’m an engineer who loves puzzles.
That Log4Shell day was rough, but it was my problem, and I got to solve it. A lot of developers don’t get that feeling anymore. Now we’re telling a model how to figure things out instead of figuring them out ourselves, and it’s nowhere near as fun. So what happens when the fun stops?
Something has to change, not only in what software people do, but in how companies see the value engineers provide. There is far more value there than knowing how to write code. Writing code was always a means to an end. There’s architecture, making sure a system is built correctly, and knowing how to solve problems.
AI is not great at creating systems. It writes excessive code that’s verbose, hard to read, and hard to maintain. It produces vulnerabilities. It’s good at finding specific ones, but not consistently, and we’re still figuring out how to automate that work. It may get better at that. But somebody still has to decide what to build and whether it’s safe to ship.
So the puzzles are still there, and they’re the kind AI is worst at. But a lot of companies are just piling more work on their engineers while they watch colleagues get fired, instead of giving them those bigger problems to own.
We still need software developers, and we still need application security people doing that work. Their day-to-day may look different. If you lead a software or security team, I’d ask a few things.
Don’t judge your engineers by how much code comes out per person. Don’t cut so deep that nobody is watching when the next Log4Shell hits. And give people systems they own, not just tickets to close. People care about what’s theirs.
The people at the keyboard bring more than the ability to type text into a model and hope the right outcome comes back. You will still need them.
When Log4Shell hit, no prompt saved my company. A person woke up, realized it was their problem, and owned the outcome. You can’t automate caring. Companies that forget that will find out the hard way.