The Hellraiser Was Supposed to Be About Tomorrow
I Thought I Had Invented the Future
I made one fairly serious miscalculation while writing The Hellraiser.
I thought I was writing about the future.
Not the distant future. I have never been particularly interested in spaceships, colonies on Mars or laboratories populated by androids in suspiciously immaculate white coats. I wanted something much more uncomfortable: science close enough to reality to make readers think, Could this actually happen?, but far enough away that, if challenged by an indignant scientist, I could smile sweetly and say, It’s fiction.
That was the plan.
So I invented an artificial intelligence capable of exploring biology on a scale no human scientist could manage and designing biological sequences that had never existed before.
In The Hellraiser, this eventually leads to Protocol Eden, an AI-designed virus created not by a cackling psychopath stroking a white cat, but by intelligent people who genuinely believe they are trying to save the world.
I was rather pleased with this.
Thriller writers are allowed to cheat. We take something that is scientifically possible, move it a few uncomfortable steps forward, remove the boring obstacles involving funding, ethics committees and people saying no, and ask the most useful question in fiction:
What if?
I thought I had moved those steps quite far enough.
Then science started following me.
I began reading about Evo 2, a generative AI model developed by researchers from Stanford, the Arc Institute and collaborators. Evo 2 was trained on almost nine trillion nucleotides, compared with roughly 300 billion for its predecessor.
Nine trillion is an impressive number, although admittedly not one that comes up often over dinner.
But that wasn’t the number that bothered me.
The number that bothered me was one.
One year.
Evo 1 had appeared only the year before.
That was the moment I stopped looking at what AI could do and started looking at how quickly the definition of what AI can do was changing.
We have a comforting habit of treating technological limitations as though they were border posts. AI can do this. It cannot do that. Everything safely beyond the barrier is labelled impossible, and we can go back to worrying about something manageable, such as our electricity bill.
Unfortunately, someone keeps moving the border.
The Most Important Thing AI May Be Changing Is Time
Evo 2 can analyse DNA, predict biological functions and generate genetic sequences, including sequences unlike those already found in nature.
Before anyone starts digging a bunker in the garden, this does not mean scientists have created Protocol Eden.
They haven’t.
Nor are laboratories casually asking computers to design the next pandemic between coffee and lunch. The developers deliberately excluded viral genomes from Evo 2’s training data as a safety measure, which is worth mentioning because conversations about AI and biology can travel from legitimate concern to Hollywood surprisingly quickly.
And the potential benefits are enormous.
This is where the doctor in me starts arguing with the thriller writer.
The thriller writer sees AI generating biology and immediately thinks, Excellent. Disaster.
The doctor thinks, Wait. What could we cure?
Systems like these may eventually help scientists understand genetic mutations, identify treatments, design medicines and investigate diseases that have resisted conventional research for years. AI might even help us design bacteriophages, viruses that attack bacteria rather than people, which could become increasingly important as antibiotic resistance grows.
So perhaps the virus isn’t the villain after all.
Annoying, isn’t it?
Villains are so much easier when they have guns, bad manners and identifiable addresses.
What fascinates me most, however, is not simply what AI may discover, but what it could do to the time required to discover it.
Science has traditionally been a long and occasionally humiliating conversation with nature. You have a brilliant idea, formulate a hypothesis, design an experiment, perform it, analyse the results and discover that nature has completely ignored your brilliant idea.
Then you start again.
Sometimes this takes days. Sometimes years. Occasionally an entire career.
Get Sophy Smythe’s stories in your inbox
Join Medium for free to get updates from this writer.
Now put increasingly powerful AI inside that process, capable of exploring enormous numbers of possibilities computationally before the physical experiment begins, and the rhythm changes.
The scientist remains essential. Biology remains stubbornly complicated. Experiments still have to work in the real world, where molecules, cells and organisms have the irritating habit of behaving like molecules, cells and organisms rather than PowerPoint slides.
But the territory that can be explored before reaching the laboratory becomes much larger.
And the next question can arrive much faster.
This presents a professional problem for people like me.
You begin writing a technological thriller about tomorrow, spend several years writing, rewriting, editing and publishing it, and discover that tomorrow has rudely arrived before the book.
There should be an authors’ union for this sort of thing.
I’m Going to Have to Write Faster
Human beings are not particularly good at understanding acceleration.
When someone tells us that something cannot be done today, we tend to hear that it will not be possible for a very long time.
Those are not the same thing.
Something can be impossible on Monday, theoretically possible on Wednesday and the subject of a start-up pitch by Friday.
That is why Evo 2 interests me less as a prediction of one particular biological breakthrough than as evidence of velocity.
When I first read about it, I did not sit back smugly and think, I predicted this.
I didn’t.
Real biology is infinitely more complicated than anything a novelist needs to squeeze between two covers, and I remain extremely grateful for that fact.
My reaction was considerably less dignified:
I’m going to have to write faster.
Yet underneath the joke lies the question that became central to The Hellraiser.
I am not particularly frightened by the Hollywood version of AI in which a machine suddenly develops consciousness, becomes furious with humanity and decides to exterminate us. Hatred seems unnecessarily theatrical.
A sufficiently capable system doesn’t need to hate us.
It needs an objective.
And human beings are perfectly capable of providing one.
In The Hellraiser, the people behind Protocol Eden believe they are using extraordinary intelligence to solve extraordinary problems. They want to save lives. They want to protect the planet. They want decisions to be rational rather than distorted by greed, politics, emotion and human self-interest.
Their intentions are not necessarily evil.
Neither is the technology.
Which is precisely what makes the whole thing interesting.
The nightmare begins when intelligence finds solutions its creators never anticipated and follows their logic farther than they intended.
That remains fiction.
For now.
But something has changed since I began writing the novel. I used to think the interesting boundary was the one between possible and impossible. Now I think it may be the distance between them.
Every better tool makes another question easier to ask. Every computational experiment can shorten the road towards a physical one. AI doesn’t merely advance alongside genetics, drug discovery, chemistry and protein engineering; it may help accelerate all of them, while their discoveries feed back into the systems doing the accelerating.
I thought I had written a novel about tomorrow.
Then tomorrow started moving.
Which leaves me with two conclusions.
As a former doctor, I find the possibilities extraordinary.
As a thriller writer, I find them mildly inconsiderate.
And as someone currently working on the next Charlie Martens novel, I have one urgent message for the scientists:
Could you possibly slow down? I’m trying to keep up.
#ArtificialIntelligence #GenerativeAI #Biotechnology #FutureOfMedicine #TheHellraiser
