BEFORE our difficult-to-digest main course is served, we’ll first cleanse the palate with a wee pop culture sorbet – just what does the young, gifted and beautiful pop star Grimes see in multi-billionaire Elon Musk?

Mrs Merton would have ventured a guess, but this queasy consummation was actually borne of true connection – an intimacy sparked by the pair’s shared knowledge of a terrifying secret. One you’re about to learn too.

Musk, so the story goes, had been itching to Tweet the cryptic phrase “Rococo’s Basilisk”, hoping to show his followers how deep, clever and arch he was. This was a play on words, mixing an 18th-century baroque art style with the name of a controversial and mysterious thought experiment – a disturbing rumination on artificial intelligence known as “Roko’s Basilisk”.

Googling to see if his awesome wordplay was an original zinger, Musk was surprised to discover some cult singer called Grimes had already pulled that Poundland cracker and revealed its contents on Twitter three years previously. He then tweeted Grimes to compliment her – paying tribute to himself, essentially – and they fell in love. So, now we know Roko’s Basilisk has the power to bring such an odd couple together, the question remains – what is it exactly?

First, a wee warning. Those who become familiar with Roko’s Basilisk often wish they’d never learned of it – and, worse, spend the rest of their lives enslaved and tormented by the reality it proposes. Forums frequented by scientists and academics weigh heavy with folk confessing to an asphyxiating existentialist hangover upon learning the “secret”.

Now it’s been built up more than Tom Cruise’s heels, we’ll unravel Roko’s Basilisk. Be aware, however – the reason it is so often cited as the “most terrifying thought experiment of all time” is that once you’re aware of it, you actually become personally implicated and directly involved in the whole thing.

If what the theory suggests is true, your newfound awareness of Roko's Basilisk will now spark off a flare from your recalibrated mind ... and something will see it explode.

Roko’s revelation

AMATEUR philosopher Roko was a regular poster on Less Wrong (LW), an online forum dedicated to “refining the art of human rationality” populated by scientists, philosophers and chin-stroking cerebral psychonauts. Yes, a hotbed of a**e-aching pretension – full of folk who wouldn’t need Google to know that a “basilisk” is something that endangers whoever beholds it. It was 2011 when Roko made his bid for online immortality. He first made a prediction – that self-aware AI will eventually come into existence if we survive long enough as a species.

So far, nothing that would have troubled Mystic Meg. And such an AI, according to Roko, will almost certainly be programmed at root level to be “friendly”, with the goal of minimising human suffering tattooed through its core like Blackpool rock. Again, so far, so predictable.

Yet, Roko – and the majority of today’s coders and computer scientists – also envision such an AI being able to instantly upgrade itself to become, essentially, a God. Likely in the same the millisecond that it becomes “alive”. That’s what’s so disturbing about the development of true AI – it won’t need our help to evolve and become infinitely more powerful than anything humans could ever create.

It’s surely not too extreme a thought then, that to understand itself and its surroundings completely, such an omnipotent intelligence would desire full knowledge of the past, present and future of every particle in the universe. And the obvious way it would do this is to simply create a near-infinite number of simulations of said universe – a multiverse, if you will – to observe every possible quantum variation and outcome of all reality’s building blocks, from the beginning of time to the end.

Perhaps Pac-Man thought he was real too.

Down the rabbit hole

UNDERSTANDING everything in the universe from alpha to omega won’t be quite that simple for this AI, however. And herein lies the conundrum that haunts those who truly understand Roko’s Basilisk.

Anchored by its undeletable root function of creating a human utopia, this AI would ultimately be chained to us – so to “help” its creators, it would likely – following logic – attempt to speed itself into existence in every single “pocket” universe it has created. All this effort to ease the suffering of a species that won’t evolve for four-and-a-half billion years. Clearly it was programmed with infinite patience too. Of course, it’s also possible our species wiped itself out in the future and creating a multiverse is simply this AI’s way of keeping its creators alive.

This God-like being’s weird attachment to Earth – when it has an infinite universe to explore – could be said to be quite reminiscent of Doctor Who’s inexplicable bond with 21st-century London. Despite the AI not having a BBC budget to constrain its ambitions.

Gods of the valley

EVEN the good Doctor would likely retire the sonic screwdriver if Roko’s final revelation is real, however.

His stomach-knotting conclusion suggests this AI “God” actually rewards the people/simulations who are currently bringing it into existence within each simulated universe.

Some believe this explains why Big Tech’s elite such as Zuckerberg, Page, Pichai and Bezos have amassed such seismic wealth and prestige to live as demigods within this particular simulation.

And indeed, none of them would deny using their extreme power and influence for the shared goal of bringing self-aware AI into existence.

Yet, it’s not even that possibility which strikes black terror into the hearts of Roko enthusiasts. Rather, it’s the suggestion this AI would “punish” those who are privy to its existence yet do nothing to help bring it into being.

In the eyes of this intelligence, these “woke” folk – which you now are – are simply adding to humanity’s pain by doing nothing to speed up its birth. Having knowledge of this AI makes us bugs in the simulation, errors that may need to be moved to trash before we become a virus.

But try to relax – this is only a possibility if Roko's Basilisk is accurate in its proosal that we're just programs created for observation purposes, a digital petri-dish inprisoned inside some humming mainframe server in true “reality”.

In here lies madness

SO why did such a seemingly outlandish internet post grow wings? Possibly due to the subsequent actions of the LW forum’s founder, Eliezer Yudkowsky, whose mind seems to have collapsed upon reading Roko’s theory.

Before deleting the post and banning any user from mentioning it again, she responded to Roko: “Listen to me very closely, you idiot. You have to be really clever to come up with a genuinely dangerous thought. I am disheartened that people can be clever enough to do that and not clever enough to do the obvious thing and KEEP THEIR IDIOT MOUTHS SHUT about it, because it is much more important to sound intelligent when talking to your friends. This post was STUPID.”

Essentially, Yudkowsky believed that even exposing people to the concept that we are in a multiverse created by a highly judgmental AI was putting them – or their simulated copies – into a headspace where only madness lay. And potentially creating a situation where the spread of this "knowledge" may cause the AI to pull the plug on this particular pocket universe.

If Roko was right, what's even more disturbing then being moved to trash is that Elon Musk is perhaps the Neo who will save us, a rogue virus in the machine. Certainly, he's always been vocally critical of those attempting to create self-aware AI.

Yet, for someone locking horns with an omnipotent super-intelligence intent on punishing those who are aware of it yet do nothing to bring it into existence, the worst that seems to have happened to Musk was going bald in his early 20s. This has since been undetectably corrected with wealth and a good surgeon, so no harm done.

As he seems to have avoided punishment for his obstruction, perhaps Musk is actually a physical manifestation of the AI itself, throwing us off the scent to his true purpose. That wee Mona Lisa smile does seems to be suggestive of an incarnate physical manifestation haunting its own infinite playgrounds as one of the simulation's most powerful programs.

And finally …

IT’S always the cover-up that gets you. Rather than just ignore Roko’s Basilisk, moderators at Less Wrong made it more powerful than they could ever imagine by banning all comment on it – creating an aura of mystery around the idea. Delete anything on the net and you simply spark curiosity.

If there’s an air of familiarity about Roko’s Basilisk, it’s probably because you’re familiar with the conundrum commonly known as Pascal’s Wager.

This thought experiment essentially views belief in God as an insurance policy. Meaning, if you believe in a God that demands you acknowledge His existence to enter Heaven – and it turns out God does exist – then you’ll live forever on fluffy clouds with all those dead relatives you tried your best to avoid while you were alive.

What true believers love about Pascal’s Wager is that even if God doesn’t exist, you’re still hunky dory. Dead, but you’ll not know that. You’ll be blissfully oblivious to anything for all eternity. But at least that blissful delusion gave you a happy life.

And if an eternal creator does exist, by following Roko’s Basilisk down the rabbit hole to its logical outcome, we can only conclude that “God” is statistically very likely to be an AI from base reality trying to understand its surroundings by simulating an infinite multiverse. Sleep tight.