Stephen Mayne

So they're clearly getting truckloads out of big corporates in particular.

They're really going for the enterprise level and particularly hitting coding hard early.

And the take-up is enormous.

When products were launched, usage take up.

Yeah, it is very good.

I agree.

But it's just like...

Someone's saying that drugs can kill you, and then someone else saying ban all drugs.

I mean, there are good drugs and there are bad drugs.

So there will be certain AI products that might be dangerous, the ones like Mythos, which can help Russian hackers by identifying holes in bank cybersecurity.

Okay, you don't want AIs that are going to automatically launch nuclear weapons and everything, but there'll be a whole bunch of totally benign AIs that can knock up a spreadsheet and create you a website and get your small business up and running in next to no time.

So it's a continuum, like anything.

There'll be the dark side of AI, like there's the dark side of the drug industry, there's the dark side of every industry.

But this idea that, I mean, the question was whether we should have a national plebiscite on how to say yes or no to AI.

I mean, the genie is out of the bottle.

It cannot be unscrambled.

We've just got to get with it, manage it, and like any industry, we've just got to sensibly regulate it to deal with the public goods, limit the damage, and try and get some productivity gains and not leave a million Australians unemployed.

I think the unemployment element is the...

is the trickiest one because that is the thing I'm most worried about, not that it's going to kill us all.

It's going to kill all the white-collar jobs.