Monday, April 22, 2024

the bleeding edge of AI

 If AI can do it, we'll be bored of it. Inherently. It treats content as a solved problem, but in reality the best it can do it make a million suggestions disconnected from the real world. That is the human part, really the whole part: WE are the thing the makes sense and understands the real world, so however end up making use of AI, we'll be the only ones who can understand what works in the given context.

All we've done is invest a million monkey middle managers to each manage their own team of a million monkeys on typewriters.

So let it ingest everything that's come before! We'll keep what works. And let it swirl around and around in its own madness it spits back out. We don't need it to live.

Tuesday, November 28, 2023

My Answers to the Colbert Questionnaire (Questionert)

  1. Best sandwich?
    • Johnny's Italian Beef, dipped w/ sweet pepper
    • A midnight Monte Cristo from the Golden Apple
    • Capicola provolone sandwich with roasted red pepper from Joe's Capri deli
    • An Alpine sub, 18" uncut
    • Turkey with avocado, dark meant only, on Dutch Crunch from the Arguello Super
    • when I'd make one from the lunch spread at my Grandma's and I'd add some of the good ranch dressing made from the packet and sour cream
  2. What's one thing you own that you really should throw out?
    • most of my clothes, honestly, and the old prie dieu
      • but no books, nope, no way, not a single one no siree bob no thank you
  3. What is the scariest animal?
    • a man with nothing to lose
  4. Apples or oranges?
    • cara cara oranges
  5. Have you ever asked someone for their autograph?
    • No, but I shook Nobuo Uematsu's hand once
  6. What do you think happens when we die?
    • The world keeps trying.
  7. Favorite action movie?
  8. Favorite smell?
    • garlic cooking in olive oil
  9. Least favorite smell?
    • that time I worked Recycling in college and we had to transport week old "compost" in an elevator
  10. Exercise: worth it?
    • yeahhhhhhh........
  11. Flat or sparkling?
    • Flat
  12. Most used app on your phone?
    • Safari, and by extension the Pocket extension, but then most of the time I don't actually go into Pocket and read it later
  13. You get one song to listen to for the rest of your life: what is it?
  14. What number am I thinking of?

  15. Describe the rest of your life in 5 words?
    • It keeps changing, so enjoy.

Wednesday, May 17, 2023

Migrating from Twitter to Newsletters

I've been following the whole collapse of Twitter via, incredibly, my inbox. 

I remember when I realized how much my mind was fraying on Twitter. The spring of 2019, I'd just started a new job and was in that phase of onboarding I'm trying to absorb a massive amount of information and context, to the point I could see the docs and code floating in the ether when I closed my eyes. And, I could feel the difference between that kind of information overload, and the one I'd get from spending my hours with twitter and feedly.

So I embarked on migration from twitter to newsletters:

  1. Cleared out my followed accounts with the delightful Marie Kondo-inspired Tokimeki Unfollow tool on Glitch.
  2. Combed through my remaining follows for as many TinyLetters and Substacks as I could find.

The best thing I did was use a "+newsletter" tag to the email I use for all my subscriptions ( will still be routed to you). Then I created a filter that automatically moves them is a separate folder, to spare my poor inbox.

A screenshot of my where all my newsletters wind up. This would have been anathema to my former inbox-zero self.

In the spirit of the internet (l'esprit de l'internet?), I figure I'd ought to share the email filter and process so anyone (even you!) can try it. 

  1. Copy or save the template file
  2. Change TKTK_YOUR_EMAIL_HERE_TKTK to, well, your email.
  3. In gmail, go to "Settings -> Filters and Blocked Addresses -> Import filters"
  4. Click "Choose File"
  5. Choose the file you just saved
  6. Click "Open file"
  7. Click "Create filters"
(And in a different kind of l'esprit de l'internet, I must note the quotes in the steps above are google's copy verbatim, and I noticed to the inconsistent case style, too.)

The result: if there's anything happening on twitter worth knowing, I'll read about it by the end of the day from at least one writer, probably more (I'm reading [checks other tab, gasps] ninety five writers on substack alone). This crystalized for me over the last few months, as I kept up with the twitter collapse, the fall of SVB, and the WGA strike. 

So that's how my media diet has been evolving. Tune in next week for my rant about how podcasts are the new TV :)

Thursday, March 30, 2023

Our simulations have been good enough for a generation

Two old stories have been rattling around in my brain lately, so I figured that's a good sign I should write this down.
A Wired story for 2010 about Madden jumping into the real world. The gist (emphasis mine):

Just before he reached the end zone, with 17 seconds remaining, Stokley cut right at 90 degrees and ran across the field. Six seconds drained off the clock before, at last, he meandered across the goal line to score the winning touchdown. For certain football fans, the excitement of a last-minute comeback now commingled with the shock of the familiar: It's hard to think of a better example of a professional athlete doing something so obviously inspired by the tactics of videogame football.

 Stokley said. "I think everybody who's played those games has done that" — run around the field for a while at the end of the game to shave a few precious seconds off the clock. Stokley said he had performed that maneuver in a videogame "probably hundreds of times" before doing it in a real NFL game. "I don't know if subconsciously it made me do it or not," he said.

And an Ars Technica piece from last year that brought this back up:

The maneuver happened at...[Martinsville Speedway]...well known for its tight, shallow-banked turns that usually require heavy braking to negotiate.

Instead of slowing down on the turn, Chastain shifted into fifth gear and gunned it, riding the outside wall and passing five cars to finish the race in fifth place.

In a post-race interview with NBC Sports, Chastain said, "I played a lot of NASCAR 2005 on the GameCube with Chad [Chastain's brother] growing up, and you can get away with it, and I never knew if it would actually work. I did that when I was 8 years old."

We call them games, like Madden NFL and NASCARthey are made for play. And they succeed at that! And in that play, players can experiment in ways that would be impossible in the real world, even unthinkable. Stokley's wall ride, to even conceive of it, requires treating a race car as borderline disposable. You have to separate yourself from the car in a way that is basically the opposite of how we are taught to drive in the "real world."

Anyway, in a nice bit of synchronicity, a line from what I'm reading right now seems relevant:

[Learning] does not consist only of knowing what we must or we can do, but also of knowing what we could do and perhaps should not do.”

Simulation, games, play in general–expand our sense of the possible. 

Monday, January 10, 2022

Spreadsheets & Slot Machines

Have you ever hit a point, playing a video game, where you get bored of the game itself? I don't mean the story gets boring, or the game getting to hard it loses fun. When the actual mechanics of the game loses its je ne sais quoi and play become rote, and, well, mechanical

This is how it happens for me: my mental model of the game abstracts away almost everything and the whole game devolves into an optimization problem. At first, I'll feel clever, like I "solved" the game. But the game hasn't ended, so what have I solved for? I've solved for the how, how to min-max my way to the fastest most efficient win condition. For me, I've come to to refer to this as seeing the spreadsheet behind the game (it's particularly strong phenomenon in strategy games, which is a bummer because I used to love strategy games).

On the other side of this spectrum would be games of purely random numbers with no place for human influence, like slot machines. Which I've noticed seems to be the core mechanic of–and I don't want to overgeneralize–about 99% of the mobile game market, god help us all.