I was reading this article about something the author calls computism, basically this idea that if we collect enough data about something, eventually we can understand it, predict it and optimise it.
And I get why that sounds attractive.
I like numbers too. If something is technical, give me the numbers. If you’re designing a bridge, please do not tell me the structure “feels right”. I would very much like some calculations involved.
But I don’t think everything works like that.
The article starts with this smart oven that was being designed to cook restaurant-level food automatically. The company had actual chefs helping develop it, except the chefs kept doing things the oven couldn’t really understand. They would taste something, look at it, change something midway, improvise. One of them literally said he needed his “chef spidey senses”. The company meanwhile wanted repeatable recipes and was worried users would develop “bad habits” by changing things themselves.
And immediately I thought: maybe I want bad habits?
Sometimes I want more sugar than I probably need.
Sometimes I want something cooked slightly more than usual.
Sometimes I make something differently just because I feel like it that day.
If you eat the statistically perfect version of the same meal every day, eventually I feel like you’re going to get bored of that too.
You could probably put enough sensors into an oven that it measures moisture, chemicals, temperature and whatever else and gets extremely good at predicting what food should taste like.
Fine.
I just don’t understand this need to optimise everything for the sake of optimising it.
Numbers are useful. They are not God.
The article talks about this happening in design too.
Clicks, engagement, conversion, productivity, A/B testing. Everything slowly becomes something we can measure, and because we can measure it, we start treating the measurement as the actual thing.
Let’s say I make a blue button because I think blue works better.
Then testing says a red button gets noticeably more clicks.
Okay.
I’m not going to ignore that because I have some sacred attachment to blue.
I would probably test red again. Maybe try another shade. Maybe combine things. Maybe ask people I trust what they think and figure out why people are reacting differently.
But I also wouldn’t immediately go:
red got more clicks, therefore red is objectively better design.
Better at what?
Getting clicked?
Looking good?
Fitting the brand?
Helping the person who uses it every day?
Those are different things.
There is that famous story in the article about Google testing 41 shades of blue to see which one got more clicks. I understand why a company would do that. When millions of people use your product, even a tiny change can matter.
I also completely understand why a designer would eventually lose their mind if every shade, border and pixel needs a spreadsheet to prove why it deserves to exist.
At some point someone has to make a judgement.
Design thinking is useful, but come on
Same thing with design thinking.
Empathise, define, ideate, prototype, test.
Useful.
Especially when you’re starting out.
It gives you somewhere to begin and stops you from immediately jumping from “I have an idea” to “this is obviously the solution”.
But actual projects are rarely that clean.
You research something and find out you were looking at the wrong problem.
You prototype something and it fails.
Someone uses it in a completely different way.
You make a random sketch and one curve suddenly becomes the thing that makes the whole product work.
An engineer tells you something cannot be manufactured like that, so now you go backwards and change half the product.
Then when everything is finished we put it nicely into a presentation and it looks like:
research → insight → ideation → solution
as if we knew exactly where we were going the entire time.
We did not.
I think frameworks are a good starting point.
They are not rules carved into stone.
A lot of the time good design seems to come from being obsessed with a problem long enough that you keep noticing things.
You need to mess things up
Making mistakes is necessary to becoming good at basically anything.
Imagine the opposite.
You never make a bad sketch.
You never pick the wrong material.
You never make a prototype that breaks.
Some system just gives you the optimal answer every time.
You might finish faster.
But what did you actually learn?
A lot of experience is probably just having made enough mistakes that when you see the beginning of the same mistake again, something in your head goes:
wait, I’ve seen this before.
That’s one reason I want to learn under experienced industrial designers.
You can teach CAD.
You can teach manufacturing methods.
You can teach ergonomics and research.
But how do you teach:
“twenty years ago I tried this exact thing, it failed horribly, my professor told me I was being an idiot, and I’ve never forgotten it”?
You don’t.
That knowledge is sitting inside the person.
Maybe a supplier told them something once.
Maybe a mould failed.
Maybe they saw users repeatedly grab a product in a weird way.
Maybe they designed something years ago and watched it come back broken.
Experienced designers seem to have little pockets of knowledge everywhere, and somehow all of those things come together when they work on one product.
I don’t think you can speedrun that.
You need to be a polymath
And I genuinely think designers should be polymaths.
Not TikTok polymaths.
Not:
Marcus Aurelius on desk, chessboard, black coffee, one fountain pen.
“I study philosophy.”
I mean genuinely being interested in a lot of things.
Games. Cooking. Machines. Art. History. Psychology. Music. Manufacturing. Whatever you actually care about.
You never know where something will come back.
You learn about one thing for no design-related reason at all, and five years later it suddenly helps you solve a completely different problem.
That’s what I mean by accumulation.
And that’s also where I think humans and AI are very different.
AI has an unbelievable amount of information.
That’s literally why I’m talking to one while writing this.
But it hasn’t accumulated a life.
It hasn’t burnt food and remembered the smell.
It hasn’t spent hours fighting with a CAD model.
It hasn’t had a supplier tell it something can’t be manufactured.
It hasn’t owned the same object for ten years and watched it slowly wear down.
It has information about all of those things.
That is useful.
But it is not the same as having lived them.
Taste is accumulated life
I think taste is partly that too.
Accumulated life.
And specifically your accumulated life.
My taste is obviously going to be different from somebody else’s.
Different place.
Different childhood.
Different culture.
Different games and movies.
Different objects around the house.
Different things we became obsessed with.
Different mistakes.
So when people talk about somebody having “good taste”, I don’t think there is one universal correct taste hiding somewhere.
Someone else’s taste can be just as developed as mine and we can disagree completely.
That’s fine.
Maybe good taste is just having an opinion that actually came from somewhere.
And I think that is why design becomes weird when we try to make something appeal equally to everyone.
You can’t.
And maybe you shouldn’t.
Make sure you’re not making anyone unnecessarily suffer, obviously. Accessibility matters. Different controls, settings, accessories and options can exist.
But the person who uses the thing the most should probably love it.
If you try to make everybody equally happy, sometimes you end up with something nobody hates and nobody particularly cares about either.
Also, sometimes the boring answer exists for a reason
There is another conversation online that goes something like:
why is nothing beautiful anymore?
And sometimes I’m sitting there thinking:
what do you want me to do bro?
Do you know how many rules an ordinary product has?
Or a building?
Fire regulations, accessibility, manufacturing, cost, structural requirements, electrical standards, maintenance, insurance, thousands of people needing to safely use the thing.
Of course some things are going to start looking similar.
I don’t think every boring building is evidence that designers have collectively forgotten beauty.
Sometimes there are just a lot of constraints.
Phones are a good example.
Everyone wants a phone that is thin, cheap, strong, waterproof, repairable and has a battery you can easily replace.
Okay.
Now make all of those perfect at once.
That is where compromise starts.
Maybe you seal the phone extremely well and now repair technicians hate opening it.
They are right to complain.
The engineer saying water resistance matters is also right.
So maybe the solution is better repair tools, replaceable seals, sensible fasteners, documentation, whatever you can reasonably do.
You talk to each other and figure it out.
That is also design.
There is not always a perfect answer
I think design is probably much more about compromise than people admit.
The designer can be wrong.
The engineer can be wrong.
Users can ask for something without realising what other problem it creates.
Research can be misleading.
The data can be completely accurate and still push you toward a product nobody really loves.
There isn’t always a secret perfect solution where every stakeholder wins.
Sometimes you just need to understand what you’re sacrificing, why you’re sacrificing it and who has to deal with the result.
And this is where experience comes back again.
Not because the experienced designer is always right.
They aren’t.
They have just seen more consequences before.
I don’t think life has “boring parts”
People always say technology should automate the boring parts of life.
I don’t fully agree.
Sometimes washing dishes is just washing dishes.
Put some music on.
Sing.
Wash the dishes.
That’s still part of your life.
Now washing an entire family’s clothes by hand every day?
Yes, please give me the washing machine.
Heavy work that destroys someone’s body?
Automate it.
Dangerous work?
Yes.
Brutally repetitive work where a machine genuinely gives someone their health or time back?
Great.
I am obviously not against technology.
I just don’t think manual labour itself is some terrible primitive thing humans need to escape.
There is nothing wrong with cooking something yourself.
Fixing something.
Cleaning.
Building.
Sanding.
Carrying.
Making things with your hands.
Effort is not automatically a design problem.
WALL-E already showed us this
Honestly WALL-E is probably the easiest example.
Great movie.
Cute robots.
Terrifying situation for the humans.
They have achieved maximum convenience.
Food comes to them.
Their chairs carry them everywhere.
Screens entertain them.
Everything is automatic.
They barely have to do anything.
And they’re basically in a prison.
The worse part is that they don’t even really know they’re in one anymore.
That’s the thing I don’t like about endless convenience.
Convenience is good.
I like washing machines.
I like power tools.
I like not standing in a bank queue for three hours.
But if eventually a system decides what I eat, what I watch, what I read, what I buy, what I make and what I write…
what exactly is left for me?
The article ends with this image of counting every petal while losing the ability to taste the cherry.
I think that’s basically what I’m worried about.
We keep treating friction, mistakes, effort, subjectivity and uncertainty like they are bugs.
Sometimes they are.
Sometimes they are also just life.
No conclusion
Use the data.
Use AI.
Use design thinking.
Use machines.
I use all of them.
Just don’t confuse any of them with the human being.
Numbers can help us decide things.
They shouldn’t automatically make the decision.
Frameworks can help us think.
They shouldn’t become religion.
AI can know an enormous amount.
It still hasn’t lived a life.
And designers can use all of those things and still eventually have to sit at a table with engineers, users, management, manufacturers and everyone else and say:
okay, I think this is the compromise we should make.
They might be wrong.
Then hopefully they learn something.
I don’t want technology to remove every mistake from life.
I don’t want it to remove every bit of effort either.
Help us where it actually helps.
Automate things that genuinely should be automated.
But leave something for us to do.
Something to choose.
Something to get wrong.
Something to learn.
Something to make.
Otherwise eventually we might optimise life so much that we stop actually participating in it.
