As we entered 2026, I kept being asked: “have you heard the rumour about DeepMind?”
The answer to life, the universe and everything is well known to be 42, calculated by Hitchhikers’s Deep Thought. What Deep Thought struggled with is the question, something which required a world model to be built: the Earth.
This is what DeepMind allegedly have made: a world model with AGI potential.
In this piece, we’ll talk about world models, why they are the route to AGI, the dangers of walking this path, and the reason why we’re screwed if we don’t.
95% of AI pilots fail to demonstrate ROI. 90% of CEOs see no productivity gains. 80% of employees report anxiety despite 88% adoption rates.
The data is screaming that AI adoption is failing and the barrier is psychological.
Meanwhile, I built a psychology-first AI enablement strategy from scratch at a FTSE 250 company. 286% ROI. The equivalent of 72 extra FTEs from 500 users. Execs telling me I changed their life.
I've since applied for over a hundred roles. Zero callbacks.
The market is hiring CIOs and ML engineers to solve what is fundamentally a human problem. And the stats show it: just 5% of firms get tangible gains, 23% sustained adoption, zero evidence that time savings translate to productivity gains.
And not just murdered. Obliterated. Erased from history. My digital self, two decades in the making, fully John Connor’d by the Zuck.
This is just the tip of Zuckberg.
In this article, I examine his far greater crime: how he unleashed one of the theoretical misaligned monstrosities of AI safety that's devastated communities, spawned political horror, and led countries to war.
Hi - I'm a former strat comms guy turned AI strategist. I've spent the past two years laser focused on adoption at a FTSE 250. Despite getting some strong results, we had the CFO blow up everyone's budget in Dec, and consequently left as the best I could do for people going forward was Copilot.
I'm typically not on social media, so come back after a long time away to see an absolute tonne of inane drivel on LinkedIn and Twitter. One thing that's really stunned me is the complete lack of awareness when it comes to risk. As such, one of the things I'm trying is to highlight risk we're seeing at the cusp of this and explain it in ways that hopefully connect with the average interested person.
Picked on Musk making Grok into Skynet in this first piece. If you enjoy it, would greatly appreciate a like or a share. If you didn't, would appreciate your feedback so I can improve for the next one. I am aware it's longer than most of the crap I'm seeing online atm, but not a fan of this whole 'give it to me bitesized' culture that's emerged in recent years.
Getting folks to care is essential in my experience. In coaching adults "what's in it for me?" is the end of my first section and forms the basis of their first prompt. Also how I cover risk - ie. "How do I not damage my credibility?". If you're asking people to break habits and processes, you've got to make them want to.
That said, the hands on approach here is great and also foundational in my experience.
The answer to life, the universe and everything is well known to be 42, calculated by Hitchhikers’s Deep Thought. What Deep Thought struggled with is the question, something which required a world model to be built: the Earth.
This is what DeepMind allegedly have made: a world model with AGI potential.
In this piece, we’ll talk about world models, why they are the route to AGI, the dangers of walking this path, and the reason why we’re screwed if we don’t.
Bring a towel.