The speed at which Agentic AI went mainstream is both amusing and satisfying. A veteran tech marketing pro, I wrote many years ago of my frustration that IT was not included in every single news story regarding business. 

Back then, I was angry because, in my view, every journalist in the land should know more and just care more about the tech which was already in every citizen’s hands and most workplaces, thanks to PCs and iPhones. In retrospect, I should have been careful about what I wished for.

The way flagship news programmes talked about Anthropic’s Mythos launch has changed my view. Never before has such ‘new tech’ been greeted with such a mix of bafflement and curiosity, bemusement and confidence. 

Mythos this, Mythos that. It was as if the hype which never in truth greeted the iPhone’s launch had come back decades too late and with an additional twist of AI doomsterism.

We must of course give a hat tip to the Anthropic marketing team’s strategy of teasing the media, and so the world, that the innovation behind Mythos, was way way more than just an AI -powered hacking team. 

The company has said that Mythos was not ready for a public launch because it is too effective at finding high-severity vulnerabilities in major operating systems. Why would this be an issue? Surely that is a good thing?“Oh No” such magical tech could result in it being abused by cybercriminals and spies. 

So instead of just letting it go wild, Anthropic made it available to a small group of the world’s biggest cybersecurity and software firms as part of “Project Glasswing”. Feeling patronised yet?

Marketing pros often say exclusivity drives value. So denying the millions who one way or another make their living from cybersecurity from accessing the code was a genius PR move. Boogeymen work well when they are masked. Less so, when unveiled and such launches are what PR traditionally does well. This was just a different tactic.

Of course, there are serious threats from the ‘malware cocktails’ which a powerful Large Language Model can create and send to the, still largely human, based cyber defence teams which keeps industries and governments up and running. There was never any way manual testing every six or twelve months, then hoping a Microsoft-inspired Patch Tuesday once a month could keep up with the pace of bad actors.

But we knew all this before Mythos. But now, thanks to Anthropic, AI’s redefining role in cybersecurity is now clear.

So finally my wish for wall-to-wall coverage on issues such as data sovereignty, decentralised systems and open source software has been granted. However, some of the public debate, especially where Government officials are non-technical bureaucrats, has been amateur hour and in some cases dangerously ill-informed. 

Who cares? For those who still see tech as obscure, optional or upsetting, Mythos is the scare tactic of a lifetime, which means it can be talked about endlessly for two or three weeks and then forgotten in this high-speed news agenda. And what next?

It is already too late for Anthropic’s rivals to be first movers. Sure, Open AI now has Daybreak and Mistral is developing a European alternative. No doubt there will be a raft of other latecomers trying to replicate the impact of this PR breakthrough – which some see as little more than a cyber scare makeover. Anthropic has what it wanted, brand recognition, first mover advantage and an IPO announcement where it looks like the category winner. 

Let’s just hope Mythos was not the ‘boy that cried wolf’ and some clever PRs, like those at Positive, can figure yet more creative ways to put a spotlight on tech’s growing importance from here on.