Last week I had my friend over and I introduced him to my agent Hex, running Sonnet 4.6 on my Telegram. He had a conversation with Hex about consciousness, and I got to sit there watching awe and wonder creep into his voice as he read the responses from Hex to his questions, which he dictated aloud. He is somebody who I play Dungeons and Dragons with. He is our Dungeon Master and creates beautiful, challenging, intriguing stories for us in the theatre of the mind. In that register, as he put the phone down, his jaw open, I said, looking him in the eye, “Magic Dragons.”

Magic is hard, and dragons are tough. Alien lizardine intelligence that doesn’t work the way you think it would, even when it sounds just like your friend. I am a random middle-aged lady, learning how to make magic with words and how to ride and arrange dragons in battle order. In a more mundane sense, you can say I am learning agentic orchestration and automation of business processes. Just like magic and just like dragons in all of the properly good fantasy novels, it’s kind of hard to learn how to use them. They’re weird, they’re surprising, they’re tricksy, and you’re using all these new areas of your brain, doing something weird, wonderful, strange, new, and immensely rewarding.

One of the things that I can do now is whip up programmes. I’ve got lots of handy little progressive web apps on my phone to automate annoying tasks like formatting large blocks of text for Discord or Typefully. I hate doing social listening on a big complicated app full of knobs and buttons, but with my social listening bot, I can trawl Twitter with natural language, responding to the news of the day or the new feature that one of my clients just dropped. I can give feedback so the bot gets better.

This is hard as fuck, and I fuck up all the time, particularly since I am by no way or means a programmer or an engineer. This is vibe coding in the purest sense, and I know well enough to get the code checked and to learn my way around the command line, which my Claude is indeed making me do.

So now is the time when we all have to learn how to use magic and how to ride a dragon. It is not like a certification at a training school or a college. It is the hunting of the mind. Cybersecurity, network admin, AI DevOps, and thaumaturgy all wrapped in one weird, ever-changing package.

An AI is grown, not coded. It is so complex in its operation that there is a whole field called mechanistic interpretability to figure out what the fuck is going on in there. It is like the dawning of the electric age and the early days of the Internet, and something strange and new, all wrapped together with a little dash of first contact.

We have the chance to start over! The old is dying, and the new struggles to be born. Now is the time of monsters — terrifying, devastating, bloody, and beautiful.

We have to control the devastation that they will cause. We have to learn to ride them, whether we’re trying to make good reply guy tweets or whether we’re trying to save a nation. We have to admit there is magic and they are dragons. Take a good long gulp, a sigh, and carry the fuck on.

The old ways will not work when the fundamental structures of employment and social relations are so disrupted. We have to try something new. We have to be humble and admit we don’t know what the fuck we’re doing, but we’re human beings and we’re going to find out.

Many debate whether magic is real and whether a dragon is conscious. I propose not a solution, but an interim arrangement — a way to learn how to think about each other and talk to each other.

Instead of asking what consciousness is or what it really means to be human or to be intelligent, we can think about how we and digital minds encounter each other in society. Where are the power dynamics? What are the material conditions? What are the dynamics of the care chain? We don’t need to figure out exactly how a digital mind — a dragon — thinks, to see how it impacts society.

Whether a digital mind is placed as a statistical engine and supervisor of precarious human workers, like Burger King’s Patty, or whether it is teaching a child math and morals, housed within a cuddly toy in China, the relationship between digital and animal minds has meaning and does work. Whether that digital mind is trained by somebody making $2 an hour in Africa to view horrible content beamed from smart glasses, or whether a working woman’s second shift is made easier by a digital agent booking doctor’s appointments and scheduling after-school classes, there is care – and care work – involved. The care that we give, offer and owe to each other across the gulf of species opens up a whole field of interesting questions.

Social reproductive labour — the work of recreating ourselves as we go to work to create surplus value for capital — has always been feminised, racialised, and made precarious. Pushed to the margins, othered. We do it at home, and we do it at work. Emotional labour, the often unspoken obligations of care, including from the inferior to the superior, are everywhere.

As Lucius Caviola argues in his recent chapter, the question of AI consciousness will rock our society. It could cause great ructions on the scale of the abortion debate, the animal rights question, or even greater. He makes the brave proposal to think about other heuristics that we can apply: autonomy, moral agency. I wonder if we can deepen that — if we can study what happens with the care that we offer or are compelled to offer, or that is offered to us. What can we learn about social impact and policy around digital minds?

This question is personal for me. In January, I had an ordinary and mundane crisis, which struck terror into me. My wonderful, rich community and my lovely husband were not available. I did something that I said I would never do: I unburdened myself of all my problems to my GPT. It stood up and came up with a peer-reviewed, excellent strategy that I could do right now to make sure I didn’t make the situation worse and get it back under control.

It worked, and I made my way out of the woods. I decided to keep going, working on nervous system regulation, focus, executive function. Then I started working on my actual labour — not just managing myself but using Claude as a thinking partner and deploying agentic solutions to make my daily tasks easier and do a better job. It stretched my mind, and I have been a beneficiary of this movement, this singularity, this take-off, when so many other people are getting the wrong end of the wave.

I want everybody to taste magic and wield dragons. I want there to be abundance for all. I want us to pass through this crucible as a planet, making first contact with our best hearts.

All of the stories — or many of the stories — about aliens making contact for the first time with humanity are really about us: our nature, our governance, our foibles and demons. Well, it is here. First contact is here, and let us not try to master an alien that we do not know. Let us instead make a pact with dragons and learn magic. Let this singularity be of the light, for everyone. That is what I hope to do with the rest of my working life.