Technology vs. politics, round 1

Radim Řehůřek politics Leave a Comment

I intend to keep this blog mostly technical. But since it’s the eve before parliamentary elections here in the Czech Republic, I feel a small politically-technical rant is in order.

The Dawn of the Computers

Let’s start with a hypothetic vision. Imagine all computers in the world suddenly stopped functioning. All hardware evaporates, all software disappears. All at once.

No more online porn, no more funny cat pictures. What did Lady Gaga wear today? Justin Bieber’s latest tweets?

Now, how many IT gurus do you reckon are there in the world, total, who could rebuild a computer from scratch? How many people could set off, using just their knowledge, brains and books, and give you back the sublime experience of reading a blog post such as this?

If you know even a little about modern computer systems, you’ve probably guessed that the answer is zero. Nobody can possibly know the entire stack that goes into making your computer display a simple blog post. You’d need to know about silicon, gates, processor architectures, peripheral devices. Quantum physics, electromagnetism, chemistry, material science, network systems, operating systems, application programming, design. A dizzying but invisible depth.

The only way to produce a modern computer is to rely on another, slightly less complex computer to do the engineering, modelling, calculating, projecting, producing. But even that one is too complex. You’ll need another, still simpler computer to manufacture it. It’s turtles… I mean computers all the way down. A bootstrapping process of humbling depth, its origins already shrouded in myths. Think about that as your phone’s signal goes into space!

Myths, you say?

OK, so what?

That prelude brings us back to politics. Because the process of doing politics (not just the end product, but the process of getting there) is driving me mad. Yes, the relationships, stakeholder relationships and compromises involved in politics are complex. But that’s nothing unique! There are many other complex systems, which nevertheless manage, with much less handwaving, to get the job done. Like online porn. Or pictures of cats.

To explicate what I mean, here’s how different problems are approached in computer science vs. politics:

Process Computer science Politics
Testing Every self-respecting computer language, framework, library, product includes rigorous testing. Unit tests, integration tests, functional tests. Entire test-driven development methodology. Ivory tower ideologies. Bouncing off the walls. Oh, tens of millions people died because of our brilliant idea? Let’s try this one instead. A hundred million? Oops. Maybe the next one will work? Ship in more ivory!
Evaluation Collecting data points, logging. Real-time data analysis and visualization. Find out true causes. Explicit performance metrics. Using training vs. holdout vs. test data sets for prediction. A/B testing. Find out what to promise once every four years. Then forget immediately. Publish a book about what went wrong, as soon as it’s safe and makes no difference.
Innovation New breakthroughs in engineering. New applications of maths. Better understanding of how the universe works and what people what. Standing on the shoulders of giants. New breakthroughs in political science (?). New ways to deepen the voters’ dependency cycle via increased bureaucracy. Standing on the shoulders of Goebbels.
Optimization A free market of ideas and products. Can build upon free/open sourced stuff with minimal costs. Both cooperation and competition of dizzying magnitude. Competition of political parties, once every four years. Upfront costs of buying the media. Pocket greasing of dizzying magnitude.
Self correction Planned maintenance. Bug reports, ticketing, JIRA. System upgrades, versioning. An occasional good old full system rewrite. The Irish Referendum. An occasional good old revolution.


Ok, this was obviously tongue in cheek. Everybody knows politics sucks, a cheap shot. The bottom line may be that experiments are not as easily repeatable in social sciences, and collecting data points takes just too long for anyone to bother.

But it’s true managing complexity in systems with too many degrees of freedom, too many unknowns and conflicting, shifting goals is on my mind a lot. It’s actually my job, pretty much. And I am genuinely curious whether someone has tried to put the democracy bullshit machine on a more solid, scientific footing. A shuddering thought: maybe they did, succeeded, and this is the best we can do.

“If it moves, tax it. If it keeps moving, use the sickle claw.”

Fun fact: I sent 25k czech crowns (= $1.3k) donation to the party I’ll be voting for tomorrow. It is a small party. Its entire campaign budget was ~$80k, which means I financed almost 2% of it myself. Makes me feel like a proper mafia boss (“kmotr” in Czech) 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *