Karen Hao

Because Chinese companies faced with this constraint on computing resources decided to develop these technologies in a different way instead of enormously complex.

scaling up these massive facilities that use the power of cities, they just use different methods to advance the same AI capabilities.

And then they opened up their models and put them online for free.

And now there are American startups that prefer to use Chinese open source models than to pay

for the models from OpenAI and Anthropic.

And so really, it's not a competition of a race to the bottom.

And it should never be, the race should never be defined as a race to the bottom of who can scale faster, more aggressively, more recklessly to achieve what end.

It should be about figuring out what is a vision of AI development that actually upholds

scientific integrity that upholds human rights, that upholds labor rights, that ultimately advances societal aims that we've already collectively agreed on in the Sustainable Development Goals, for example, and then have a race to the top with

the competition of ideas.

So, you know, China didn't actually play the US game.

They didn't take the same model of scale, scale, scale.

And they literally subverted it entirely by taking this more free and open approach.

And so now the question is, what kind of vision is the UK going to put forth?

What kind of vision is Europe going to put forth?

There is an alternative AI vision that we could rally behind that would become

a one that actually demonstrates that we can use, develop, build this technology in a way that doesn't undermine people having dignified lives.

I do think it actually should be central to politics, but in the sense that every single person, not politicians, every single person who is a citizen of the world should be talking about this technology.

Because if we do not collectively participate in its governance, we are going to allow these companies to empire build their way.

Into a deeply anti-democratic world order.