A primer on Magic: The Gathering

For a long time in Magic: The Gathering strategy discussions, there were considered 3 types of decks: Aggro, which tries to spend its resources on threats as fast as possible to end the game quickly, Control, which spends its resources efficiently answering threats, trying to spend fewer resources on the answers than were spent on the threats until it has run its opponents out of resources and can end the game however it wants, and Combo, which avoids using resources until it can deploy them all on a single turn, ending the game. Typically, Aggro beats Combo, as it ends the game before it has accumulated enough resources to win, Combo beats Control, as the answers the control deck is playing don’t apply to what the Combo deck is doing, while the Control deck stalls the Aggro deck out of cards with board wipes and card draw spells.

A pivotal deck in the development of the game was called “The Rock and His Millions”, named after Dwayne “The Rock” Johnson, which defined a new archetype: Midrange. Midrange decks are part of the way between Aggro and Control - they attempt to deploy their resources proactively, but they still try to generate advantages by taking away their opponents’ resources - a commonly good midrange card is a creature (which can attack and win you the game if it isn’t answered) which also makes the opponent discard a card. This is therefore a “2-for-1” - you’ve played one card and gotten two effects cards out of it. (the creature which can win you the game and the card your opponent has to discard) The Rock was a sweet deck, and some of the decks that play its signature card “Pernicious Deed” are some of my favorite - “Gifts Rock” is a staple of the Extended format that was around when I was getting into the game, and remains my favorite format ever. In fact it’s too sweet. Magic players fell in love with value, and the game hasn’t been the same ever since. The game developers realized that players enjoy playing proactive threats that generate value, and so they kept on printing ever increasingly strong value engines. This meant that the goal of looking at every new card individually is replaced with one central method of evaluation: do I get the most cards out of my mana?

Credit Card Points

There’s a subculture obsessed with credit card rewards points. Websites will show you how to get awesome first class vacations for doing nothing outside your normal spending. All you have to do is open the right credit cards, and soon you too can fly on Emirates Suites for your next amazing vacation. But the central point of this subculture is the valuation - how much are the particular points worth as compared to other ones. Entire websites are built around the idea of finding you the “best redemption” (being the most valuable), and telling you how many cents your points can be worth. This is to the point where airlines’ entire profit centers come from selling these points to the credit card companies, and those companies keep coming up with new cards for people; evaluated on one metric: how do I get the most points from my spend?

Gym Culture

There’s this idea, and backlash of using science, or bro science or whatever to find the “optimal” way to train. You need to do the “Big 4” lifts because that hits the most muscles or you need to use perfect form with full range of motion, since that’s how you avoid getting hurt. And hell, it’s all a pretty good rule of thumb for beginners, just like you usually want to look for “2-for-1s” in Magic.
But it’s boring as shit, people who are slightly past beginners easily start to major in the minors and the backlash is nearly as boring – “JUST TRAIN HARDER BRO”. Still at a certain point you realize every routine, sport and exercise has come to be evaluated on one metric: how do I gain the most muscle from my time?

Reddit is Midrange

I think this is kind of my central thesis, and why I wanted to write this. People on Twitter were asking “What is Reddit”, with some scathing takes about books, movies etc. I think the central idea of “Reddit” is that it’s Midrange - in jobs, restaurants, in how you choose what to buy (“buy it for life”), in where you vacation, in how you participate in your hobbies or do your job; you should spend your resources in the most efficient way possible and you’re a chump if you do anything else.

Don’t wait in a long line to see that tourist attraction in the city you’re only visiting for one day in your life, it’s not worth it! You need to grind Leetcode to get a job at a Big-N company so you get a salary worth of your time at work. You have limited resources, be that time or money, and you need to spend those resources to generate the best “ROI”. You see news stories about Las Vegas rooms being empty, and people talk about how the experience “doesn’t live up to the cost”. A beer with your friends is going to fuck up your sleep score. We meal plan, because it “hits the sweet spot of time, money and macros”.

Middle-class striverdom taken to the heat death of the universe.

We’re spending so much time planning, we aren’t living.

A Postscript on AI

I think the LLMs turn this up to 11 because they make this stuff so much cheaper and easier; you don’t even have to plan as much. You can just ask what the best X is, or have it review and adjust your resume; so why shouldn’t you? Hell, I had an LLM edit this blog post! It isn’t worth asking another person to read it, but a free API call will make it marginally better so why not? Free Value!

So who knows? For right now I’m making fun of SWE Bench because it’s a terrible metric, only caring about Python Code. But maybe someone will finally write the one true eval, and that single metric will allow the machine god of gradient descent to finally produce the perfectly smooth Shoggoth that has all of humanity on the true global maximum of life.

But until then, stop optimizing and smell the roses.