Apr 2, 2022

A cleanup fund

"When you start a tech company, you have to allocate a portion of your stock essentially to a forecasting of possible ways that things could go wrong and you have to actually allocate resources into that.

Knowing there's going to be problems and you actually have to
allocate a portion of your stock to that. So that now there's a skin in the game process.

There's an actual stakeholder that's in your cap table as a company that is resources that are devoted to anticipating negative consequences and also ameliorating them when they show up.

As opposed to, if I did that, I'd be increasing my cost as a company and, if I'm Facebook and I do that and TikTok doesn't do that, then I'm just losing."

...

This is a transcript from episode 49 of a podcast called 'Your Undivided Attention' with Tristan Harris.

These posts are meant to be conversational. Feel free to leave a comment on Twitter https://twitter.com/joshpitzalis/status/1510080044433305600

Shoutout to Jasmin Sessler for the thumbnail.

Josh

More from Josh

Aug 31, 2025

Every clever shortcut is tomorrow’s bottleneck

When AI model get faster and better, any implementation details are just a form of premature optimisation. They constrain AI to our understanding of a domain and limit its ability to operate independently. Tell the system what to look for and what doing that well means, and then get out of the way. When building AI products, the goal is to be able to swap out all of your…

Aug 27, 2025

The Bitter Lesson

If you're building anything that involves AI, then you have to know about the bitter lesson. The bitter lesson is an article by Rich Sutton from March 13, 2019. This sounds like ancient AI lore but is only six years old. Rich Sutton, a pioneer of reinforcement learning and a Turing Award winner, wrote a short essay saying that 70 years of AI has taught him and the best quote…

Aug 24, 2025

Using Evaluations Offensively

Most people test what AI can do. Ankur Goyal suggests testing what it should be able to do. Typically, as product developers, we build a feature then write some tests to protect it. When tests fail in the future, you know you've broken something important and need to fix it before pushing any changes. This is true of conventional code as it is for evals. Goyal, who heads…

Dec 22, 2024

Spontaneous Preparation

Started playing Dungeons & Dragons again and Im easing back into being a dungeon master, something I enjoyed as a child. There are two big challenges: Preparation time and wielding narrative tension The reality of playing DnD as an adult is that time is scarce. And preping for a game as a Dungeon master takes up a lot of it. Robert Peake, over at https://www.spontaneousdm.com…

Nov 25, 2024

Giving Good Thread

I recently started playing Dungeons & Dragons again after about 25 years. I used to play it as a kid, but now that I’ve picked it up again as adult, I’m with the idea that you can sit around and talk, yet this structured, elaborate immersive experience unfolds. Now we have a regular tuesday game online, and a one or two in-person games each month, and I’m starting to warm up…

A little more from Josh.

More from Josh

Every clever shortcut is tomorrow’s bottleneck

The Bitter Lesson

Using Evaluations Offensively

Spontaneous Preparation

Giving Good Thread

More from Josh

Every clever shortcut is tomorrow’s bottleneck

The Bitter Lesson

Using Evaluations Offensively

Spontaneous Preparation

Giving Good Thread