This week's token stream

Things I learned

There are more deaths from alcohol each year in the US than all illicit drugs combined — Charles Fain Lehman via the Ezra Klein Show

Worth your time

How Density Zones can solve America’s housing Crisis — Agglomerations

A Character.ai character’s (potential) role in a teenager’s suicide— NYT.

Musings

One thing I notice is that super early stage companies have to do impossible things to make it to the next day… which sometimes leads to them being bad at assessing which super impossible things they can do and which ones they can’t. I think political progressives can be the same way. They have to by nature believe that wholesale change is possible but this blinds them to the things that can’t be changed.

With apologies to the boffins around the world who know more about this than I do, I think AGI arrived with GPT-4. Everything else is just a continuation on that theme.

The thing that’s great about Duolingo is that they’ve figure out how to turn aggressive gamification strategies towards a neutral to positive end

“A good science fiction story should be able to predict not the automobile but the traffic jam.” — Frederik Pohl

Great advice from This is Not Advice:

When starting out, it’s easy to spend time on your strengths and ignore your weaknesses until much later. In fact the startup world often fetishizes a founder's particular strength. That is to say, if you know a particular founder is design-oriented, you expect to see some of the most beautifully designed things from them. Despite this, if you know you are capable of doing something, it can be beneficial to focus on the other things. Simply put: start with the stuff you don’t know you can do. [Ed.: I remember a startup in which we, foolishly, passed on investing that had made such fast progress. When I asked them how they'd advanced so quickly they explained that they'd exclusively focused on things they were unsure would work.]

I’d add Alex Komoroske and Simon Wilson to the meme creator list.

2024-10-24