I recently wrote about AI and productivity, and how data from the 2024 Accelerate State of DevOps Report from DORA shows that widespread AI adoption in the software industry is contributing to a real and meaningful decline in software delivery performance. Approximately 76% of developers use AI tools in daily tasks such as coding, debugging, and documentation.
This week I posted a silly and whimsical post on Bluesky, which seemed to resonate with a lot of you out there, and it reminded me of a section in the DORA report about valuable work.
AI has always promised to “help people spend more time doing valuable work” by “automating the manual, repetitive, toilsome tasks” so that software developers can be “free to use their time on ‘something better.’” Despite this, the report states that “individuals are reporting a decrease in the amount of time they spend doing valuable work as AI adoption increases”. The maths isn’t mathsing.
What is valuable work, actually?
I have observed a growing trend of developers focussing solely on the tools used to make software, rather than what the software actually does. Many people are sharing their new apps on social media, attempting to provide context for their creations by listing the databases, runtimes, frameworks, UI libraries and AI code generation tools they used. But what does your app actually do? What problems does it solve? Tell me about the value you just created!
Valuable work and meaning is not derived from what AI makes us (apparently) faster at: generating code. Meaning and value in software development is actually created through the impact of building things that makes human lives better, or easier, or slightly less bad. Now, it can be argued that much of the work in the technology industry in 2025 is not centred on making things better for humans whatsoever, but that’s a discussion for another day.
What’s becoming clear is that the mass adoption of AI is shifting the focus away from human-centred software solutions that provide meaningful value, and is reducing the entire industry to just the tools at its disposal. Just generate the code, bro. Just ship one more app, bro.
The new kitchen metaphor
If I employ someone to build a new kitchen for me, I really don’t care what drills, hammers, nails, or sandpaper they use to get the job done. I just want a valuable end result: a fancy-looking and functional kitchen that makes good use of the available space, enabling me to cook delicious food in a delightful and comfortable environment. Ultimately, a great kitchen is created with vision, creativity, and by solving existing problems the old kitchen presented. The same is true for software.
Value in software development cannot be determined by how many lines of code you can bash out in any working day, and especially whether or not you are using AI to do so. Real value is delivered through vision, creativity, experimentation, and using human brains to solve human-centred problems. The report backs this up:
[T]here is also an art and empathy underlying a great product. This might be difficult to believe for people who think everything is a problem to be resolved through computation, but certain elements of product development, such as creativity or user experience design, may still (or forever) heavily rely on human intuition and expertise.
And what’s even more interesting, is that while seemingly “high-performing teams and organizations use AI”, the report finds that “products don’t seem to benefit”. We’re all just churning out AI generated code, moving those tickets, making meaningless graphs go up, but to what end? Real software is actually not moving forward. We're all just cranking out the same broken software with the same stupid bugs. In fact, as I was editing this post whilst sitting with a fresh batch of black hair dye on my head at the hairdressers, one wrong tap of a button on my phone deleted half of the article. I attempted to use the three-finger gesture to bring up the undo button on my iPhone, which, not surprisingly, did not restore what was deleted. I ended up having to find a deploy preview of this post (that I fortunately deployed before I left the house), copy and paste half of the article back into the CMS, and reformat the headings, being very careful to not make a single wrong move.
All the new kitchens and silly add-ons are shit. There’s no value in that.
Tools do not determine value created
The tools someone uses to build a kitchen are only as good as the skills of the person using them. A skilled craftsperson can probably use any old tool and produce a great result that holds up for years to come. A less experienced craftsperson who doesn’t understand the fundamental concepts of space, structure and value-based utility might be able to make a single kitchen cabinet look good to the untrained eye, only for the shelves to be at the wrong height so I can't store my stuff. The same can be said for software.
There’s nothing wrong with being inexperienced; we all have to start from somewhere. But we can’t rely on tools as a shortcut to gain valuable experience. Experience takes time to develop, and your tools are only as good as your fundamental knowledge and skills. If you skip the knowledge and skills part, and if you fail to learn about what you’re doing and the implications of how you’re doing it and the human value you have the potential to deliver, then you have little hope of building human value into your software. Because for the most part, humans use software. Andreas Møller said it better in They lied to you. Building software is really hard:
The true value of a software engineer is in our ability to analyze problems as well as design and implement creative solutions. To get good at these skills you need to understand not just the tools at your disposal but also the technologies you are building on top of. If you don’t understand how an application works then you have no chance of fixing its bugs and issues.
In the report, “respondents also reported expectations that AI will have net-negative impacts on their careers, the environment, and society, as a whole”. Whilst AI has empowered anyone to build and ship web and mobile apps, the tangible negative impact of the soulless, valueless software released on a daily basis across the industry cannot be underestimated.
And no, I do not want to use your new AI tool to summarise an email (?) or summarise a document (?) or summarise a meeting (?) or anything else. I want to use my brain so I can comprehend and learn and connect with what is in front of me.
Productivity ≠ value
Returning to the topic of the previous article, I want to touch again on the topic of productivity. Productivity as a measure of value is extremely misleading. The very notion of productivity has been conjured up by Big Capitalism™️ to keep us busy, and to misdirect our attention so that we forget to question the broken system in which we have no choice but to participate. All of this prevents us from real growth, and moves us further and further away from the pursuit of real value.
You could argue there is space for both approaches in software development. Build the hard, compelling stuff using your human brain, and use AI to generate the code for some of the more boring parts of your apps. Take form validation, for example, scaffolding out a new project, or setting up all the boilerplate to make some API calls. But those boring parts are only boring because you wrote the code to do the same thing before! You already learned how to do it without AI.
But unfortunately, too many of us are getting sucked into the productivity hype cycle and engaging in daily conversations with energy-sapping computer machines that vomit out thousands of lines of code based on probability and existing mistakes that the Large Language Models themselves are trained on. It's absurd. We have a whole new cohort of inexperienced kitchen builders who have invested in the latest must-have drills, hammers, nails and sandpaper, but have no idea how to build real value into what they are using those tools to create. And so we're seeing an influx of infinite inferior kitchens that offer no human value. They may look good when you walk into the room, but will inevitably fall to pieces as soon as you put the kettle on to make a cup of tea.
The data speaks for itself. Vibe coders are reporting outages and critical security vulnerabilities in their apps, losing months of work that didn’t use version control, and the inability to truly learn new things.
The world is already cooked, and yet, we’re cooking it more
What’s more, there’s the environmental impacts of AI, which are only just beginning to emerge. An MIT article titled Explained: Generative AI’s environmental impact outlines how the rapid expansion of generative AI presents significant sustainability challenges, including electricity and water overuse, hardware-related emissions, and increasing pressure on power grids.
The world is already cooked. And yet, we’re cooking it more — literally, and figuratively — by shipping insecure software into the void that we have no idea how to debug, scale, or extend. In the not-so-distant future, LLMs will be trained purely on LLM-generated software, and the world will eat itself.
I challenge you to find the value in that.