The Ultimate Dream

· br11k's blog

Coding is not solved. What problem solving actually is, why participation beats winning, and why writing specs is the ultimate accountability.

The Ultimate Dream

Hey MPJ, thank you for everything so far.

I am hoping to see both of us navigate this complicated, and complex world. It is weird to see that you are confused no less than myself. You have been a guiding light for me so many years ago. So I guess I have to step in and share something I feel needs to be said out loud.

Coding is not solved.

Hell, software engineering is not even coding, although many think it is.
You see, I have been on this quest for 3 months, to find meaning in my work.
I dived so deep into my OCD-fueled, ADHD-powered brain,
I believe I can now explain what problem solving is,
and how we can even measure it outside of those silly IQ tests.

In a practical sense.

What we do first is understand the World around us.
And it is essential we do this before it changes.
You see, we are not living in a static scenario
with some interactive elements, like in a video game.

The game itself is evolving
Rules update. Win condition is changing.
How do you keep winning?
That's a wrong question to answer.

People chase success all the time and they burn out.
Like you did. Like me, 150 days into unemployment.
This is not my first. It would not be my last.
So what should we ask instead?

Think about it.

I'll give you the hint.
The winning is not the point.
It was never the point. Not even in esports.
Only few teams could take the pedestal anyway.

The point is participation.

So before we even get to the point of AI agents "solving coding" and all that,
we should ask this question.
Or, rather, a set of questions.

Do they have skin in the game?
Do they understand the World similar to how we do?
Are they participating?
Or simply doing the matrix multiplications?

You probably want to point out that multiplications,
or any sort of computation is able to solve problems.
And you wouldn't be wrong.
I believe our brains are solving hardest differential equations
on a budget of 20 Watt-Hours.
This is even more impressive than a cluster of H100s.

If you think about it,
H100 was created by this very limited thing called brain.
Can only hold at most 4 things in the working area.
Addicted to scrolling, nicotine, dopamine.
How can it advance the World? Understand it?
Comprehend? Make it all Coherent?

And yet...

That's what it just did.
It wrote those lines you are reading.
So addictive. Almost if you read next word, you learn something useful.
Something important. Something that enlighten you and pull you out of depression.
And you wouldn't be wrong.

And yes, you, Pangram lovers, you can go and check this article pretending to be a comment for AI.
I love wasting your free credits on my walls of text.
It's so funny.
But I digress.

The second part of engineering is building intent.
We do it all the time.
But it is never durable. Nobody wrote a product spec.
It's all in your head.
System works, and yet nobody knows why.

A terrible situation to be in.

I hope Spotify maintains specs.
I hope Google have detailed product specs on their products.
I hope every other company keep track on what they are building and why.
Because I am yet to meet a single team, or even a person outside of Joel Spolsky who enlightened me since 2015.

I know why his idea failed.
I know why Intentional Software failed.
I know why nobody wants to pay 4-digit sum for requirement tracing software suite.
I know why nobody writes specs.

Because once you wrote something down, you are now the subject to accountability.

Apparently, being accountable is a human thing.
Coincidentally, it is also most paid role in the World.
You don't want some random person without Understanding suddenly become accountable about this tiny World you are living in.
You want the best one. The best money can buy.

The one you can trust.

I never questioned authority. See where it got me.
Unemployed immigrant in another country with nothing to pay my October rent.
Well, I can sell my 5090...
But you get the point. And I don't want to make it political and ugly.

The third thing we do is verifying intent with how the reality actually works.

The so called "curse" of reading Pull Request and understanding it.
So many people never understood the point of why we even need code reviews.
Do you care about the World you are living in?
Or simply trusting someone else to architect it for you?

And I wish I am simply exaggerating.

Code is everywhere. On your left hand.
Under your fingers. Click-clack-tip-tap.
Believe it or not, even flushing toilet is impossible without code.
When you start thinking in systems.
And I went as hard as I could, I understood what exactly code is.

Code is Data. And Data is Code.

Lisp, yes. Emacs users rejoice.
Clojure adepts ahoy.
And your product specs about how your tiny World must behave is simply that.
Code.

Wait but technically, we can't easily transform code to data and vice versa.
And you are correct.
Data becomes code only if there is a compiler, algorithms and runtime.
Our World runs on laws of physics we discovered so far, and if we can digitize time, you can see at every time quant the state is changing continuously.

Physics are source Code.
Spacetime and physical dynamics are Runtime.
Current Universe's physical State is Data.
State transitions is the Execution.

So you can formulate World as a computation engine with State, Code, Runtime.
And we are still trying to find a compact description of runtime.
We are literally reverse-engineering reality itself.
That is the ultimate goal.

"This is the story of the ultimate dream, man's dream. A dream of escaping from the confines of his own planet. A dream of freedom within his own universe, a dream of reaching the stars." (c) X: Beyond the Frontier.

https://www.youtube.com/watch?v=d_n5WEA3tDI

And we're only getting started.

last updated: