Cal Newport

So I look at this chart.

What are they charting?

Oh, they're charting tweets per day that they've detected tweets about AI doing things that you didn't want it to do.

And I said, huh.

So when does this line start going up?

the week that open claw was released to the public and everyone just started building their own bad agents and then tweeting about how bad they were.

And you know what word was not mentioned in that article?

Open claw.

And even though the examples they were giving... So they just said, scheming just started rising.

I guess AI is becoming sentient.

And then I looked at the biggest spike.

I was like, well, this day in February on this chart had the biggest spike.

It was like, oh, there was this one tweet about OpenClaw erasing someone's emails, and then it got retweeted.

It went super viral.

I was like, okay, great.

You just...

The real headline of this article, letting people write their own agents leads to terrible agents.

That's it.

And also, Opus is not doing that.

The stupid OpenClaw program you wrote that's prompting Opus and then taking action on your computer based on what it says is deleting your files.