Build a Better Basilisk
The robots are coming to murder you, or not.
You’re going to hate me for this one.
Do you play The Game?
The Game is dead simple.
If you think about The Game, you’ve lost The Game, and have to immediately1 tell someone.
The catch is, that by learning about The Game, you are immediately entered into The Game, and you can never stop playing The Game.
So you’re now a player. Sorry.
Anyway, there’s a pants-shittingly terrifying spin on the game called Roko’s Basilisk.
Roko’s Basilisk is a hypothetical future insane robot intelligence which will punish any and all humans who are aware of the possibility of its emergence but who do not actively work to bring it about.
Basically, it’s The Game. Because you know about Roko’s Basilisk, you’re now at risk of a digital replica of your sentience being eternally tortured by said Basilisk unless you drop everything and start developing cyber-Mao.
Before you start panicking and explaining the concept of gulags to a LLM in the hope you’ll be made the commendant, don’t worry. I did some deep research, and find it highly unlikely that we’re on the cusp of the emergence of some AI super-bastard.
Now as I’ve ably demonstrated, we’re not at risk of Chat GPT setting itself up as a digital Satan to purgator us until the heat death of the universe.
But we need to take a leaf out of Roko’s book and start building a better basilisk.
Not because we’ll be tortured, but because if AI starts doing the boring bits of your job really quickly and your clients find out you’re not actively working to make that part of your process, they’ll quite likely bin you off for someone who will.
Ok.
You’ve scrolled back up.
You’ve checked.
This is still Unmemorable Title. This isn’t a LinkedIn Gimpfluencer shittering on about AI.
But here’s the thing. Underneath all the noise, the gimps have a point.
Clients are living in an AI world. And while they don’t like it churning out guff, they do like it doing things quickly.
If you’re training an AI to do the fun, clever bit of your job, you’re replacing yourself. And not particularly well.
If you’re training it - like we’re trying - to do all of the information gathering and grunt work - you’re building a better basilisk.
If an AI can put a dossier on my desk of a client’s current positioning compared to their competitors, and an idea of what best practice is for their industry, it’s saved me two days and primed me to start doing the creative work.
I’m still the expert. I’m still valued. The customer’s still getting great work. And they’re getting it faster.
That’s what I’m talking about.
That’s how you win the game.
Something mint - the power of laziness
So why do customers want us to get a computer to do the leg work?
Why do we want that computer to do all the collation and sorting so we don’t have to start our market or competitor or customer research with a blank sheet of paper?
Because humans are lazy. And doing things is hard.
People want to default to the easier option, even if that’s not the most optimal outcome for our wallets, mental health, or general wellbeing.
Take something as unedifying as paying your damn taxes. Sure, champagne socialists like yours truly pretend we’re happy that handing a big wodge of wonga over to the government is our civic duty for schools and hospitals and desperate people in need of shelter - but really, we’d rather be spending that cash on booze, meals, and tickets to football matches.
Not paying tax is a better proposition than paying tax. So how do you get people to pay up without going down the “we’ll lock you up, unless your name is Bezos, you freeloading bastard” route?
Appeal to laziness.
It takes minutes to pay, months to successfully avoid paying. So says the DVLA.
Nobody said an appeal to emotion has to be an appeal to the better angels of our nature.
Sometimes a resigned shrug is just as effective as a fist pump.
Or like 46 hours later in a newsletter. Samantha, I’ve just lost the game. Not sorry.