Ray TraylorMar 10
It's been bothering me why people I-know-who-know-better won't admit the real risk of AI coding on important codebases.

I'm not talking about AI evangelists, non-technical, or even devs without security/uptime ownership.

This is even outside social media, so you can't blame the algo.
Not in the context of arguing, so not rhetorical tricks.
Not even directed at me, so no identity performance.

What the hell is going on?
Show threadRay TraylorMar 10
In my little corner of Threads, four people admit it cleanly - @zellski @jwynia @__ericelliott @780.mf

What do they have in common? Two are near retirement, one is building a solution, one dgaf. Then there's me - the one who says what you don't want to hear.

Deductive reasoning easily shows employment isn't our concern.

What now?
Show threadRay TraylorMar 10
The employees who know better won't say anything. That's not quite prisoner's dilemma, more like a collective action problem -

Everyone privately knows the risk but thinks everyone else doesn't, so no one speaks.

What does that remind you of? Chernobyl.
Classic in dictator, top-down systems like corporations.

How to fix it before disaster?
Show threadRay Traylor
The solution is a preference cascade.

1. People privately believe one thing
2. They publicly express the other thing because social pressure
3. Once someone speaks honestly, everyone switches

It looks like opinions changed overnight but they were just hidden.
People lie about their true beliefs.

That's whats happening with AI and AI coding.
That's what I want to fix.

Even if I don't, at least I-know-that-you-know.
Mar 10 at 6:37pmWeb