AI, at your own risk

So Mike, let's kick this off with a Reuters exclusive.

And this one is a biggie.

We've got eyes on Anthropic's IPO prospectus.

It's unveiling its sweeping AI vision, but also those surging costs.

But I think the big headline that everyone's going to grab their attention to today is that it is planning to caution incoming investors about maybe the existential threats and catastrophic risks to humanity.

I mean, this is a big, big warning from a company that plans to profit off this tech.

It's sort of a case of, this is our product.

It could end all humanity, but please buy our shares.

It is a bit like that.

It does feel like a prospectus warning about such dangers.

And okay, these dangers have been on the front pages of all the newspapers and news sites for the past month or so.

But it's there in black and white in the prospectus that not just whether these AI models can develop bioweapons or anything of that scale, but the way in which they collaborate with each other, the way in which they bypass security checks and kill switches, the way in which they even open up blackmail of other agents and other users.

So this is This is pretty alarming stuff.

And you might then wonder, with a company warning of this, why it's still planning to go ahead with the IPO at all, at least before some broader discussion about the regulation and the ways that the guardrails, as it's continuously called, around these things actually is in place.

Now, as it stands, it appears that Anthropic will go ahead with the IPO shortly after the midterms, or at least sometime after the midterms.

And we know that OpenAI, its major rival, has postponed its IPO until next year.

So there is a question here about how soon this can go ahead.

They're aiming for a $2 trillion valuation, which would be the biggest IPO in history.

And there have been, in the background, discussions between Dario Amadei, the Anthropic CEO, and Donald Trump.

another last night, by all accounts, previously on Sunday.

So you have to wonder what that discussion is about.

Clearly, Trump has made it clear he doesn't want new regulation on these models.

So what else is afoot? Does the US government want to somehow take part in the IPO? What is it looking for? And what is Anthropic feeding back to it, more importantly? Now, there was debate about Anthropic being involved in US defense and the US kind of excluding it from those contracts.

But it's a lot of wait and see.

In the meantime, OpenAI has shelved plans for its latest ChatGPT model, this Astra 6, ChatGPT 6.