Ok, I will fully cop that it is kind of cool to pick up my phone and find a ton of activity on an old thread that I kind of liked.

Ye gods.

Yesterday I spoke at a high school in Calgary, giving the students a quick rundown of my creative career and things I wish I'd known when I was their age.

Here's a summary of key points I covered:

• My career is nothing like what I envisioned when I was young, but key parts of it shaped my future.

I mildly regret that while I am a cynical bullshitter, I am insufficiently malicious to truly take advantage of this marvelous age of cynical bullshit.

I would be AMAZING at planning theatre.

On a purely boring, mundane level, my current frustration is people turning to LLMs to handle planning tasks which are genuinely easy to do if you are actually planning, but perhaps benefit from the help when you are engaged in planning theatre.

Especially because my inner nerd really does want to celebrate when the tech can do things that are neat or interesting, but every time that happens, it is overwhelmed by the wave of crap that washes up alongside it.

All of which makes talking about any part of this mess a real pain.

Put differently, it’s an unbounded system, and unbounded systems are what tend to produce the biggest disasters.

At this point, the actual utility of AI and the destructive pattern of behavior around AI are almost entirely decoupled. The tech could just as easily be tupperware.

A more subtle danger of lottery planning is that is has no end trigger. Once you decide it is a good investment, it will NEVER STOP being a good investment, because the reward remains just around the corner. It is a behavior of addiction shaped from math.

There’s a strategy to this, in the same way that buying lots of lottery tickets is technically an investment strategy. If there IS some magical application that allows for exponential growth, forcing use increases the likelihood of finding it by sheer chance.

It also justifies infinite investment

To get value out of LLMs (which is possible) requires the ability to distinguish what isn’t valuable, and the model of “use at all costs” is selecting AGAINST the ability to do so. Bad use is rewarded, avoiding bad use is punished.

It is not necessarily easy to understand what data is actually useful. In theory, being able to make that distinction would be useful in an AI centric environment, but in practice, I doubt it will be, because no one who judges success needs to understand it.

The one curious thing I’ve seen about those areas where LLM have actually been good at things have been areas where there is an abundance of high quality data.

This is interesting, but it makes me all the more nervous about its application to areas WITHOUT such abundance, or without understanding.

AI agents fundamentally change tech jobs from writing code or pushing pixels to asking an AI to do it and reviewing the output.

This is a fundamentally different skill set and sadly I agree with Nikhyl that many companies will take the shortcut of layoffs and rehiring than retraining their staff.

Apparently men protecting children is highly suspicious behavior, if you ask the right nutjobs.

Ok, I am going to have to listen to this, as it’s adjacent to one of my great, unresolved areas of interest- early American committees of correspondence.

Pre revolutionary information science is super cool stuff.

I am a PhD sociologist who works outside academia and I approve this message.

A) This is amazing
B) I kind of want to see great matches captured as 3d sculptures, like they did with chess matches.

The video game stupid broke through my bubble, and I am now aware of something that only makes me stupider as I think about it, so I need to go stare at static for a while.

Wait.

Fuck.

I have no idea where to find static these days.

Game in a Minute: Thief’s Market | Gameosity Thief's Market from Dave Chalker and Pickpocket Games is a treasure-swiping game of dice drafting and manipulation, with just enough thievery to keep things interesting! Find out more in this week's G...

Thief’s Market from Dave Chalker and Pickpocket Games is a treasure-swiping game of dice drafting and manipulation, with just enough thievery to keep things interesting! Find out more in this week’s Game in a Minute! www.gameosity.com/2026/04/19/g...

Don’t just let it be terrible, LEAN INTO it being terrible, because no one’s going to see it.

I only bring this up because someone mentioned just this week that this trick is useful, but I *still* 30 years in, when stumped, write this on the top of the page.

Should we trust our audience more?

(I have no good answer)

Palantir recently published a 22 point manifesto.

TL;DR DEI is bad, fascism is good.

FOLKS!!!!

@naomialderman.bsky.social has bought Zombies, Run!

Yes, the old app a bunch of us used to all exercise together to and do our mini events is back in her hands!

Small thing, though. The banks apparently were empty 👀

So there's a couple of things to do to help out (mini thread)

Learn.

One thing that's worth repeating: There is *zero* evidence that "substack helps you get readers". The only thing that we've ever seen that helps grow an audience is other authors mentioning your blog or newsletter, which is not the platform's doing. Substack just takes credit for that!

When you use AI to do your writing, you are telling your audience: “I deserve your attention, but you do not deserve my effort.”

Look, for my money the absolute game changer technologies right now are batteries and biosciences, not statistically modeling a mid conversation, but you do you.

Honoured to be in great company at the IGDN Indie Groundbreaker Awards! What Should We Have Tomorrow? Has been nominated for Game of the Year and Graphic Design, and owes a lot to @lostwaysclub.itch.io who was responsible for the graphic design and game consulting. www.igdnonline.com/groundbreakers

This doesn't happen when software is designed with a user in mind.