The IT industry is only ever as good as the software it produces. There has always been great software, and there has also always been a lot of bad software.
The people writing the great software did so quietly, behind the scenes while the people writing the bad software hung out on sites like stack exchange, asking questions and ultimately attempting to answer them.
Most were one trick ponies. Hence the annoyingly common answer to “how do I do this in language X” being “No idea, but why use X” and if you’re lucky, “Here’s how to do it in Y”. that’s all they could contribute because that’s all they knew.
The people writing bad software got themselves into the news. Failing systems make for good stories. The people writing great software didn’t get noticed. Their systems just worked, and when things work no-one praises the engineers. Instead they worship the rich man who got richer on the back of their success.
This is the training set. The hobby projects, the keen amateurs, the one off prototypes. Most of the good stuff never makes it into the public domain, so the AIs have to learn from what does. Even the curated AIs have to rely on the trainers knowing what good looks like, and many of them don’t. Just take a look at the Claude leak – that tells a story all on its own.
But even so, there were always great engineers doing great things, and they still are. Not for the limelight or even for a thank you. Just for the satisfaction of a difficult job done well.And for the joy of working with other skilled people.
Unfortunately, the news cycle still pushes the bad ones to the fore. The noisy “ten apps in one day” braggers and the naively arrogant “coding is a solved problem” hackers.
Even in your own company, the team working well is the one you know little about. Why would you? They make no noise and leave no footprints to follow. They never appear in the reports unless you look closely enough to see them lurking in the shadows of successful business outcomes.
The inevitable result? Those who don’t understand the importance of great engineering industrialise the wrong thing. They automate the mass production of bad software in the belief that more is more.
When everything gets worse and things keep breaking they’ll turn en-masse to the engineers and say “this is your fault”. And the hackers who called themselves engineers will exclaim “this job’s too hard”. They’ll blame the machine, the AI and the management. Anything to avoid admitting they were faking it until they never quite made it.
Meanwhile, behind the scenes, the great engineers will carry on producing great software that just works. Quietly and unnoticed, while a rich guy in a suit gets his face on the cover of Time magazine.
People will ask, how is this done? How do we make this better. And the loud brash bluffers will go on LinkedIn, preach at conferences and write blog posts and books about how to do it. They’ll still be wrong.
Turns out, actions speak louder than words. Unfortunately very few people are listening.