21/08/2026
I keep seeing the term “AI slop” thrown around, as though the fact that AI was involved in making something automatically makes it worthless.
I think we’ve been here before.
There was a time when you drilled a hole with a hand drill. You held it steady with one hand and turned a crank with the other.
Then came the electric drill.
Suddenly the same person could work faster, drill more accurately and get considerably more done in a day.
Did the electric drill destroy craftsmanship?
No. It gave the craftsperson a better tool.
Guitars are another good example.
For generations they were shaped almost entirely by hand. Today, many excellent builders use CAD and CNC machinery to make incredibly accurate cuts and repeat them consistently.
The CNC didn’t suddenly become a luthier.
Somebody who understands guitars still has to design the instrument, choose the materials, understand construction, make decisions, assemble it, finish it, set it up and know when something isn’t right.
The machine simply gives that person leverage.
And that’s pretty much how I came to AI.
I’d already been programming for years and had always written little programs and scripts to make repetitive jobs easier.
Scripts for managing computers and servers. Python programs for repetitive tasks. Photoshop actions. File-management utilities. Anything where I found myself thinking, “There has to be a better way of doing this.”
Then in 2023, the company I was working for introduced us to ChatGPT.
One day I wondered:
Can this thing write Bash?
It could.
Then I tried Python.
That changed things considerably.
Instead of spending time typing every line and occasionally hunting for some stupid syntax error I’d introduced, I could describe what I needed, provide the environment, language, requirements and constraints, and get a starting point almost immediately.
But I still needed to understand what the code was doing.
And that distinction matters.
Over time I started applying the same idea to other parts of my work.
Research.
Copywriting.
SEO.
Website development.
Testing.
Image optimisation.
Security and vulnerability checks.
Analytics.
Eventually I realised that work which once required several different specialists over several weeks could increasingly be coordinated by one experienced person using AI as a collection of specialised tools.
Then came agents.
Instead of one general-purpose AI doing everything, I could build specialised agents for different jobs and have them work together.
The result isn’t that I suddenly stopped working.
It’s that I can now accomplish in a week what would once have taken a team considerably longer.
And I think that’s where a lot of the conversation about AI goes wrong.
AI doesn’t replace knowledge. It gives knowledge leverage.
Give somebody who knows nothing about design an AI image generator and they can produce bad design much faster.
Give somebody who doesn’t understand marketing an AI copywriter and they can produce meaningless marketing copy by the truckload.
Give somebody who can’t program an AI coding tool and they can potentially build something they don’t understand and can’t maintain.
That’s AI slop.
But give powerful tools to someone who understands the underlying craft and they can use those tools to do things that previously required far more time, money and people.
That’s not particularly new.
We’ve been doing it throughout human history.
The hand drill became the electric drill.
Hand shaping was joined by CNC.
The calculator didn’t destroy mathematics.
Photoshop didn’t destroy graphic design.
Digital recording didn’t destroy musicianship.
And AI isn’t automatically destroying creativity or craftsmanship either.
There are absolutely things happening with AI that deserve criticism.
I’m a musician, so I’m particularly uncomfortable with companies training systems on artists’ work without permission and then offering tools capable of imitating those artists. Deepfakes, scams, misinformation and malicious uses of AI are real problems too.
But those are arguments about how a technology is developed and used.
They aren’t arguments against the technology itself.
A tool can be used badly.
It can be used lazily.
It can be used dishonestly.
It can also be used to help scientists analyse enormous amounts of data, help researchers find patterns humans might miss, help small businesses compete with much larger companies, help programmers build better software and allow one person with knowledge and an idea to create something that once required an entire team.
So when I see something dismissed simply because “AI was used”, I think we’re asking the wrong question.
Don’t ask:
“Was AI used?”
Ask:
“Is it any good?”
Does it solve a problem?
Is it accurate?
Is it useful?
Was there thought behind it?
Did somebody knowledgeable make the important decisions?
Because ultimately, I don’t care whether somebody drilled the hole with a hand drill or an electric one.
I care whether they drilled the hole in the right place.
Rod.
