That's the best thing anyone has said all day as far as I'm concerned

When there's a gold rush, don't find gold, sell the tools

It’s classic MLM dynamics

• The money makes itself! It’s impossible to not make money with this system!
• You get to be the boss!
• If you’re not making money, you must be doing it wrong — my (paid) training course can help
• Heaps of unsold product rotting in garage/warehouse
• Religious-coded language
• Requires infinite growth to stay profitable for all current players
• FOMO, “getting in on the ground floor”
• Mid-levels taking huge financial risks to onboard more down-levels

Yeah like the show Devs, the AI was hidden away from the public so only the creators could use it.

That’s nonsense. They would be selling the poor to the AI. Think the Matrix, but it’s powering a stock market machine that makes money for Bezos.

Bingo.

It's a new god for the poors.

Efficient AI hypothesis

It's supposed to be stocks trading, not an AWS bill.

I mean, that happens with CloudWatch all the time. It's the most plausible part about this.

Even a vibe coded wouldn’t be that stupid… I hope.

look at various vibecode subreddits.

at least some of them are

Yes, they would.

run the first attempt with all their wallet info in production? Even a vibe coded wouldn’t be that stupid

Oh the things I have seen cryptobros on twitter do, you would not believe

I hope they would - it's like natural selection but in the fiscal context

The first three drafts of my favourite test program (a program to drive servos to point a heliostat mirror at mirrorLattitude, mirrorLongitude, mirrorAltitude in altitude and azimuth to direct the reflected ray at targetLattitude, targetLongitude, targetAltitude) wouldn't run.

It's very rare for any AI to use real modules for the language (even when I say "please use library modules that exist in CPAN") and they nearly never know how to reflect a ray

No, he didn't invent anything and doesn't exist

But think about how quickly he did it! /s

I think the last line is.

It's actually pretty funny to think about other AI scrapers ingesting this nonsense into the training data for future models, too, where the last line isn't enough to get the model to discard the earlier false text.