Schools measure compliance:
1. Did you show up?
2. Did you follow instructions?
3. Did you fill out the worksheet?
Real learning measures understanding:
1. Can you explain it?
2. Can you apply it?
3. Can you teach it?
We’re optimizing for the wrong metrics.
Build kids.
How did Claude forget that it can't scare problem solvers when it created the AI for them?
Don’t they know that problem solvers just walk away from a rusted tool to solve the problem?
I told my daughter today that Claude code is misbehaving.
She told me that maybe my prompting wasn’t right today.
What she doesn’t know is that the Claude 4.8 release is the cause, not the prompt.
How come 4.7 suddenly became -4.7 after the release of Claude 4.8?
When asking Claude to write claude.md from the discussion,
why is it not referencing the Claude Code creator's prior guide on how to create it?
Or research Anthropic docs first on what to include or exclude, or query the user via their tool.
@bcherny
Drive the AI slowly like you were driving a car for the first time.
Don't work on multiple projects at once.
Trash ideas that don't excite you.
Work on one meaningful project.
Drive relevant traffic.
Optimize for conversions.
Retain existing customers.
And then start scaling.
Talking to smarter folks than me, I'm convinced many of the AI folks in my timeline are full of shit.
Nobody is "running 20 agents over night" and building stuff for actual users. Maybe some are building internal tools or disposable software. Maybe.
But building software people like using? That doesn't get hacked on day one or blow up after the 3rd user? Nope.
I don't even understand what that's supposed to look like. Do you work out a 57 pages document that perfectly describes what you want to build and then summon 14 agents and have them run wild for 6 hours? And what comes out on the other end isn't a broken pile of shit?
Nope. Not buying it.
PS: it may also be that I have an IQ of 82 and can't figure it out.