Brian

But the proponents of governed intelligence argue that these filters actually make the AI better as a commercial product.

That is a huge chunk of time.

You don't spend 40 percent of your resources on something unless you think it adds serious value to the end product.

That's a very fair point.

And honestly, looking at the sources, it seems like the public generally agrees with this approach.

I was looking at the future of free speech survey in our stack.

Right.

And across the board, support for things like

AI generated deep fakes of politicians is below 40 percent.

In the US, it's down at just 21 percent.

So there is a clear democratic mandate for some level of restriction.

People generally don't want a totally Wild West where literally anything goes.

They don't see a seatbelt at all.

So let's pivot to the case for raw intelligence.

The core question here seems to be who decides what is safe?

And it's not always obvious when it's happening, is it?

The sources talk a lot about soft moderation.

It's not just the AI throwing up a red flag and saying, I can't answer that.

How so?

And we actually have hard data on this.