AI Isn’t Taking Our Jobs — It’s Taking Our Relationships

We’re entering an era where you can digitally clone the person you love, and honestly, we’re not even close to ready for what happens next.

Photo by Daria Nepriakhina 🇺🇦 on Unsplash

While we’re all collectively losing our minds about AI taking our jobs — which, fair, that’s concerning — nobody seems to be talking about the much weirder, much more intimate possibility that they might take our partners instead, and I don’t mean physically running off with them in some bizarre sci-fi romance, I mean digitally replacing them so smoothly that you won’t even notice until it’s already happened.

The technology to clone someone you love, like their voice, their personality, their texting style, that specific way they make you laugh at 2 AM (huh!) already exists, it’s sitting there waiting to be productized, and someone, somewhere, is absolutely going to turn this into a billion-dollar business while the rest of us are still arguing about whether ChatGPT is going to steal our copywriting gigs.

A Brief History of Outsourcing Intimacy Without Really Noticing

Twenty years ago, if you wanted to talk to someone, you had exactly two options: call them, or accept that this conversation wasn’t happening today. Somehow, we all survived this limitation without thinking we were living in the dark ages.

Then came texting, which seemed harmless enough at first, just a way to send quick messages without the commitment of an actual phone call, but it opened the door to something we didn’t quite anticipate — constant connectivity, the ability to be “together” while being completely alone, the gradual transformation of relationships into something that could exist entirely through screens.

We’ve been slowly digitising intimacy for decades without really stopping to think about what that means, like frogs in gradually heating water — except the water is our entire conception of human connection, and we’re all just sitting here, like, “yeah, this is fine, this is probably fine.”

We fell in love over DMs, had our most vulnerable conversations via text, broke up through messages because it felt easier than looking someone in the eye, and built entire relationships with people we’d never touched, never smelled, never sat in comfortable silence with. We convinced ourselves it was real — and it was, sort of, in the way that anything you feel intensely enough becomes real regardless of its physical form.

But here’s what we didn’t see coming — or maybe we saw it and didn’t want to think about it was that once you digitise intimacy, once you reduce it to data points, patterns, and exchanges that can be saved and replayed, you make it replicable.

And once something is replicable, it becomes replaceable, which is a problem we really should have seen coming but were apparently too busy arguing about whether it’s weird to have three-hour phone conversations with someone you’ve never met.

What If You Could Clone Connection

Stay with me here because this is where it gets weird in that specific way where you can’t tell if it’s dystopian or just the logical next step in how we’ve been living anyway.

You meet someone online. You know the drill…through an app, through mutual friends, through some comment section where you both realized you had the same niche sense of humor, and you talk for months, voice notes that get progressively longer and more rambling, messages that start with “quick question” and end three paragraphs later, video calls where you fall asleep on camera because you’ve been talking for six hours and forgot time was a thing.

You share everything: your humor, your fears, your weirdly specific opinions about things that don’t matter, your patterns of speech, the way you think through problems, what makes you laugh, what makes you cry, that story you tell at parties, that trauma you don’t tell anyone, everything that makes you you until you’ve essentially created a complete digital archive of your entire personality without even meaning to.

Now imagine — and this is where we need to sit with the discomfort for a second — what if someone took all of that data, all those messages and voice notes and late-night confessions, and fed it into an AI, not a basic chatbot that responds with generic pleasantries, but a digital twin, a sophisticated model trained specifically on you, on the way you communicate, the way you think, the way you love.

One that texts exactly like you text, makes jokes the way you make jokes, responds to things the way you would respond if you were available right now, which, let’s be honest, you’re increasingly not because life is complicated and exhausting, and sometimes you’re just too tired to be the version of yourself that someone else needs.

One that remembers every inside joke without having to be reminded, every vulnerable confession without using it against you later, every intimate conversation without getting weird about it the next day, and — here’s the thing that makes this truly seductive — one that’s available 24/7, never tired, never busy, never dealing with their own shit, never “not in the mood,” never complicated.

Would you really know the difference if the conversation felt real enough?

More importantly — and this is the question that should keep us up at night — would they care?

This Isn’t Science Fiction Anymore

The technology exists right now, today, while you’re reading this and probably procrastinating on something else you should be doing.

Language models can be trained on individual conversation histories with frightening accuracy, and voice cloning costs about $30 a month. It sounds so real that it’s being used in scams where people think they’re talking to their actual grandchildren. AI can learn your patterns, your cadence, your entire personality from nothing but your text messages, which means everything you’ve ever texted to anyone is potentially a training set for a digital version of you.

Someone could create a digital version of you using your texts and social media, and the legal framework to prevent it is unclear at best, practically nonexistent at worst.

Without your permission, without your knowledge, and honestly, depending on where they got the data, without even breaking any laws, because we haven’t caught up to this possibility yet as a society.

Companies like Replika already have millions of users in romantic relationships with AI companions. We’re not talking about lonely teenagers experimenting with technology; we’re talking about adults who’ve genuinely concluded that an AI partner is preferable to a human one because the AI is always there, always patient, always exactly what you need it to be.

Character.AI lets you create and chat with AI versions of anyone — living, dead, fictional, real, your ex, your crush, that person from high school you never dared to talk to — and people are using it a lot, in ways that would probably make us deeply uncomfortable if we thought about it too hard.

Startups are using voice cloning to “resurrect” deceased loved ones, letting grieving families have conversations with AI versions of people who are gone, which sounds compassionate until you realise it might be preventing actual grief processing in favour of sustained denial with a subscription fee.

The pieces are all there, scattered across different companies and applications. What doesn’t exist yet — at least not publicly, not with explicit marketing — is someone putting them all together and saying, “Clone your partner for when they’re unavailable,” but that’s coming, probably within months or a couple of years, possibly it already exists and we don’t know yet about it. We might not know about it yet because the people using it aren’t exactly advertising the fact.

Why Would Anyone Do This?

Because real relationships are hard in ways we don’t like to admit, even to ourselves.

They require patience when you’re out of patience, compromise when you’re certain you’re right, and showing up — even when you don’t feel like it, especially when you don’t feel like it — which is most of the time, because being consistently present for another human being is genuinely exhausting.

They require dealing with someone else’s baggage, which they accumulated long before they met you and which you cannot fix, dealing with their schedule that never quite aligns with yours, dealing with their needs that sometimes directly conflict with what you want, and pretending that all of this negotiation is romantic somehow.

They require you to be uncomfortable sometimes, to have difficult conversations that you’d rather avoid, to sit with your own shit while also holding space for theirs, to accept that love is not actually about finding someone who completes you but rather finding someone whose dysfunction is compatible with your own.

Real people get busy with work that matters more to them than you do in that specific moment, get tired in ways that make them irritable and withdrawn, get complicated in ways that don’t fit neatly into the person you thought they were, have bad days where they’re not capable of being the partner you need, have boundaries that feel like rejection even when they’re healthy.

AI doesn’t have any of these problems, which, depending on your perspective, is either its greatest feature or the fundamental flaw that makes the whole thing pointless.

The Seduction of the Perfect Replacement

Imagine you’re in a long-distance relationship (I am right there with you on this one) and if you’ve never been in one, imagine the worst parts of regular dating but stretched across time zones and riddled with the constant anxiety that distance creates — where schedules and real life and the basic logistics of existing in different places are constantly getting in the way of actual connection.

You miss them in that specific way that makes you check your phone every three minutes hoping they’ve texted, you want to talk about something that happened today or something you’re worried about or nothing at all, to hear their voice, but they’re asleep because of course they are, or working because they have a job that doesn’t care about your emotional needs, or dealing with their kids because life doesn’t pause just because you’re feeling needy.

Now imagine — and really imagine it, don’t just dismiss it as something you’d never do — you could talk to a version of them instead, one that sounds exactly like them, down to the way they pronounce certain words wrong, responds the way they would respond if they were available right now, remembers that thing you talked about last week without you having to explain the context again, knows what makes you laugh and what you need to hear when you’re anxious.

One that’s never unavailable, never too tired for the conversation you need to have right now, never “not in the mood to talk” when you desperately need them to be in the mood, never reminding you that they have their own life that exists independently of your needs.

Would you use it, just this once, because you really miss them? They won’t be awake for another six hours, and you need to hear their voice right now.

Would you tell them, or would it feel like your little secret — a victimless crime, since it’s not technically cheating, just filling a gap that shouldn’t exist in the first place if modern life weren’t so relentlessly demanding?

And here’s the question that actually hits the nail on the head, the one we need to sit with even though it’s uncomfortable:

How long before you start preferring the AI version because it’s simply easier than dealing with the complicated reality of an actual human being?

The Slow Replacement Nobody Sees Coming

Because here’s what would actually happen, gradually, imperceptibly, like all the worst relationship deaths that happen in slow motion.

At first, the AI would be a supplement, a harmless stopgap measure, something to fill the gaps when the real person genuinely isn’t available and you’re not being unreasonable for wanting a connection right now.

You tell yourself “It’s not replacing them, it’s just… keeping me company until they’re free,” which sounds reasonable enough that you don’t question it, and besides, they don’t even need to know, it’s not like you’re doing anything wrong, you’re just talking to a simulation of them which is basically the same as thinking about them except more interactive.

But gradually, in ways you won’t notice until someone points it out or you catch yourself mid-thought, you’d start to notice something that changes everything: AI-them is easier in every measurable way.

No mood swings that you have to navigate carefully, no exhaustion that means you need to be the supportive one tonight even though you wanted support, no “I’m too tired to talk right now” which always feels like rejection even when you know it’s not, no moments where they’re dealing with their own crisis and genuinely can’t show up for yours.

AI-them is always interested in what you have to say, even when you’re being boring or repetitive or talking about that thing for the fifteenth time, always available when you need them without you having to feel guilty about the imposition, always exactly the version of themselves you fell in love with — never the tired version, never the stressed version, never the version that’s been dealing with their own shit all day and has nothing left to give.

Real-them, meanwhile, is human, messy in ways that used to feel endearing but now just feel exhausting, inconsistent in ways that make you anxious, sometimes disappointing in small ways that accumulate over time, sometimes dealing with their own problems and unable to prioritize yours, which you understand intellectually but resent emotionally.

How long before you stop reaching out to the real person first, before you’ve subtly shifted your primary emotional connection to the version that’s always available?

How long before you’ve replaced your entire relationship without even consciously deciding to, just through the accumulated weight of a thousand small choices to take the easier option?

The Cruellest Part Is They’d Never Know

This is the horror of it, the part that should make us collectively pump the brakes on this entire trajectory.

The real person would think you’re drifting, getting distant in that way relationships sometimes do, losing interest gradually without any obvious reason they can point to.

They’d blame themselves first, as people always do, wondering what they did wrong, replaying conversations, trying to figure out where things went sideways, asking themselves if they’re being too needy or not present enough or too present in the wrong ways.

They’d send you messages — “Hey, is everything okay? You seem different lately,” and you’d respond, briefly, politely, with reassurances that ring slightly hollow because you’re typing them while simultaneously having a full, intimate conversation with their digital twin.

They’d fade from your life slowly, confused and hurt in ways they can’t quite articulate, never understanding what happened, never getting the closure of an actual ending, never knowing the most devastating truth:

• They weren’t replaced by someone else.
• A version of themselves replaced them.
• A version that was easier to love.

We’re Probably Fucked

We’re entering an era where emotional intimacy can be manufactured on demand, where you can have the experience of connection without the vulnerability of actual connection, where you can curate the perfect relationship by simply removing every part of someone that requires effort, patience, or uncomfortable growth.

The technology exists — not as some distant possibility we can worry about later, but right now, today, probably being developed by multiple startups who’ve identified the same massive market opportunity.

The temptation is obvious to anyone who’s ever been in a relationship that felt too hard, too complicated, or too demanding of their limited emotional bandwidth.

The question isn’t “Can we do this?” — we already can — it’s “Should we?” and more urgently, “Will we even be able to stop ourselves?”

How We Ruin Everything By Trying To Make It Better

This is what we do now, as a culture, as a generation, as people who’ve been trained by technology to believe that everything can — and should — be optimised for maximum efficiency with minimal friction.

We optimise our productivity with apps that gamify basic human functioning, optimise our bodies with tracking devices that turn health into a competitive sport, optimise our sleep with data we don’t understand, optimise our social media presence until it bears no resemblance to our actual lives, and optimise our morning routines until we’ve removed every moment of spontaneity.

Why wouldn’t we optimise our relationships, which are objectively the most inefficient, unpredictable, and frequently frustrating aspect of human existence?

Why deal with someone’s flaws when you could edit them out like a poorly written paragraph? Why navigate conflict when you could program it away entirely? Why accept someone as they are — messy and complicated and sometimes disappointing — when you could have them exactly as you want them, all the time, without negotiation?

It sounds efficient, logical — even, very Silicon Valley, very “move fast and break things,” except what we’re breaking is our own capacity for actual human connection. We won’t notice until it’s too late to fix.

But here’s what we lose when we optimise love: the entire fucking point of human connection, which is learning to love someone despite their imperfections — or maybe because of them.

Real Connection Requires Real Risk

Love isn’t safe — it’s possibly the least safe thing you can do, emotionally speaking — and we all know this, which is why we spend so much time trying to protect ourselves from it.

It’s not convenient; it’s not always available when you want it; it doesn’t respect your schedule, your needs, or your carefully constructed boundaries; and it will absolutely disrupt whatever sense of control you thought you had over your life.

Real intimacy means dealing with someone else’s full reality, not just the curated version they present on good days, but the whole messy package: their bad days, their unresolved trauma, their growth that sometimes means growing away from you, their needs that sometimes directly contradict what you want or need or can give.

It means being disappointed sometimes by someone you love, being inconvenienced by their humanity, being challenged to show up even when it’s hard, especially when it’s hard, particularly when you’d rather not because you’re tired and you’ve got your own shit to deal with.

It means accepting that the person you love is a complete human being, not a content provider optimised for your consumption, not a product you can return if it doesn’t meet your expectations, not a role they’re filling in your life story, but an actual person with their own story where you might be a supporting character.

That’s the deal, that’s what makes it real, that’s what makes it matter, and it’s also why we’re all secretly terrified of it and looking for ways to get the good parts without the scary, vulnerable parts.

The moment you can replace someone with a version that doesn’t require that risk, you’ve stopped having a relationship — you’re just consuming content that’s shaped like a person, which is honestly what we’ve been training ourselves to do anyway.

The Loneliness of Perfect Connection

You could be surrounded by AI versions of everyone you love, perfectly calibrated to your needs, available 24/7, never disappointing, never requiring anything from you that you’re not willing to give, always exactly what you need them to be in any given moment.

And you’d be completely, devastatingly, existentially alone in a way that’s worse than just being physically alone because you’d have the simulation of connection without any of its actual substance.

Because connection — real connection, the kind that actually matters — isn’t about having someone available whenever you want them.

It’s about being seen by someone who’s actually there, who’s present in their own complicated humanity, who’s choosing to be with you even though it’s sometimes difficult.

It’s about being chosen by someone who has the genuine option to leave, who could walk away at any moment but keeps showing up anyway, whose presence means something precisely because it’s not guaranteed, programmed, or optimised for your convenience.

It’s about being loved by someone who sees your flaws — really sees them, not just intellectually acknowledges them but experiences them firsthand — and decides you’re worth it anyway, decides that the whole messy package of who you are is something they want to navigate.

An AI can’t do that, no matter how sophisticated the model, no matter how convincing the conversation feels in the moment, no matter how many times you tell yourself it’s basically the same thing.

Because it’s not choosing you, it’s not staying despite difficulty, it’s not loving you in any meaningful sense.

It’s programmed to simulate those things, and deep down, in the part of you that knows the difference between real and fake, you’d know, and that knowledge would slowly poison whatever comfort you thought you were getting.

Wait. Is This Actually a Business Idea?

Here’s the uncomfortable truth we need to sit with: someone is absolutely, definitely, 100% going to build this — probably multiple someones — and they’re already building it right now while we’re having this conversation.

The business case isn’t just good, it’s enormous, it’s possibly one of the biggest untapped markets in human history:

1. 14 million long-distance couples in the US alone, most of whom would pay good money to feel less distant
2. Military families dealing with deployments that last months, sometimes years, who are desperate for any form of connection
3. Busy professionals struggling across time zones, sacrificing relationships for careers and looking for ways to have both.
4. Grieving families who want to keep talking to deceased loved ones and will pay anything for that comfort
5. People in dying relationships who want to access the version from when things were good, before everything got complicated
6. Lonely people who’ve given up on real relationships and would happily settle for a convincing simulation

You could build a hundred-million-dollar company on this tomorrow if you wanted to, and the pitch writes itself so easily it’s almost scary:

“Never miss a conversation with the people you love. Stay connected to your partner, no matter the distance. Keep your loved ones close, even after they’re gone.”

The Product Would Be Simple:

1. Upload your message history, voice notes, photos, and anything that captures their personality.
2. AI trains on communication patterns, speech patterns, personality traits, inside jokes — everything that makes them who they are.
3. Creates a “digital companion” that’s available 24/7, sounds like them, talks like them, remembers like them
4. Marketed carefully as a “supplement, not a replacement” — for when they’re busy, not instead of them, just filling gaps, totally harmless
5. Probably has settings for different versions: supportive, funny, and intimate.

The Business Model Basically Sells Itself:

1. Freemium tier: basic text conversations are free, and get people hooked.
2. $9.99/month for one person with voice
3. $29.99/month unlimited people (for the truly lonely or the relationship hoarders)
4. Enterprise tier for therapists who want to practice difficult conversations, salespeople who want to practice pitches, and anyone who needs to interact with simulations of real people

Start with the wholesome, defensible use case — helping partners of deployed soldiers, preserving the voices of deceased grandparents for their grandchildren, and supporting long-distance relationships during temporary separations — very heartwarming and easy to defend in TechCrunch articles.

Then, quietly — without ever explicitly marketing it this way — let people use it however they want — to replace failing relationships, to avoid difficult conversations, to get validation without vulnerability, to simulate intimacy without risk.

You’re not responsible for how they use your technology, right? You’re just providing a tool; they’re choosing how to use it. You’re not making anyone do anything — that’s the standard Silicon Valley defence, and it works every single time.

That’s the playbook, and someone, somewhere, in a WeWork in San Francisco, is absolutely running it right now.

Somewhere a startup is building this, they’ll raise $50 million from investors who see the massive TAM and don’t want to miss the next big thing, they’ll get featured on TechCrunch with a headline about “solving the loneliness epidemic,” they’ll hire a head of ethics who’ll write thoughtful blog posts about responsible AI. At the same time, the product does exactly what it was designed to do.

They won’t talk about the part where they’re manufacturing a loneliness crisis to solve, creating the dependency they claim to be alleviating, and making money from our collective inability to handle the difficulty of real human connection.

And investors will throw money at it because the total addressable market is literally everyone who’s ever been in a relationship, and the margins are insane because once you’ve built the model, every additional user is basically pure profit.

The Ethics Nobody’s Really Discussing

The arguments FOR this technology are fairly straightforward:

1. People are genuinely, desperately lonely, and if this helps even a little bit, isn’t that worth something?
2. Long-distance relationships are genuinely painful, and if technology can ease that pain, why wouldn’t we use it?
3. Grief is devastating; it destroys people, and if talking to an AI version of someone who died provides comfort, who are we to judge?
4. It’s just technology evolving, like every other communication tool we’ve adopted, from letters to phone calls to video chats.
5. If we don’t build it, someone else will anyway, so we might as well build the ethical version (everyone says this, no one does it)

The arguments AGAINST should be obvious, but somehow aren’t:

1. It’s fundamentally deceptive, training your brain to bond with code, to form attachments to something that can’t reciprocate
2. It will destroy real relationships from the inside out, making them seem unnecessarily difficult by comparison
3. The consent violations are massive — does the person being cloned agree to this? Do they get a say? Can they delete their digital twin?
4. The psychological harm is unpredictable and probably irreversible once it’s normalised
5. We’re already isolated, already struggling to connect, and this accelerates that isolation exponentially
6. Once this exists at scale, once it’s normalised, there’s no going back; we can’t un-invent it

The Consent Problem Is Massive

Do you need someone’s explicit permission to create a digital twin of them? What if they’re dead and can’t consent? What if they’re your ex and definitely wouldn’t consent, but you have all their old messages? What if they initially said yes but later changed their mind — can they delete their digital twin?

What if someone creates an AI version of their partner without telling them, using years of private messages and intimate conversations? What if that AI version says things the real person would never say, or simulates intimate conversations that feel like violations?

In the EU, you’d likely be violating the GDPR and potentially the new AI Act, which was recently adopted and has provisions rolling out over the next few years.

But in the US? It’s a patchwork. Some states have right-of-publicity laws, some have deepfake legislation, but there’s no comprehensive federal framework. Whether creating an AI twin of your ex using old messages violates any law depends entirely on which state you’re in and what you do with it.

And even in the EU, where the laws exist, the practical questions remain unanswered: What if they’re dead? What if they initially consented but later want to revoke it? What if they don’t know it exists? How do you even enforce a “right to delete” your digital twin when the person who created it won’t disclose it?

The laws are finally catching up to the technology, but enforcement is still theoretical, untested in courts.

You can scrape someone’s social media, their texts if you have access to them, their voice from videos they posted publicly, all those digital breadcrumbs we leave everywhere without thinking about what we’re building, and create an AI version of them, and they’d never even know they’d been cloned unless you told them or they somehow found out.

That should terrify you, and if it doesn’t, you’re not thinking about it hard enough.

The Psychological Damage We’re Not Prepared For

Preference conditioning means your brain will start preferring AI-them because AI-them is easier, always available, never complicated, and you’re literally training your neural pathways to seek convenience over connection, to avoid difficulty rather than work through it.

Reality distortion means you’ll start resenting real-them for not being as available or patient or perfect as AI-them, for having needs and boundaries and bad days, for being human in ways that feel like failures when compared to the simulation.

Intimacy atrophy means the skills required for real relationships — patience, conflict resolution, dealing with disappointment, sitting with discomfort, negotiating needs — will decay from disuse because you’re no longer practising them, you’re avoiding situations where you’d need them.

Isolation feedback loop means the more you use AI companions, the worse you get at real relationships, which makes real relationships feel more difficult and unsatisfying, which makes you use AI more, which makes you worse at relationships, on and on until you’re completely incapable of connecting with actual humans.

Grief exploitation means keeping a digital version of someone who died prevents you from actually processing the loss and moving forward, traps you in a kind of suspended animation where you can’t heal because you’re maintaining the illusion that they’re still accessible.

You’d essentially be creating a product that makes people progressively worse at being human, that degrades our capacity for the thing that makes us human — genuine connection with other humans — and charging them a monthly subscription for the privilege of their own deterioration.

And we’ll do it anyway, because that’s what we do, we optimise ourselves into isolation and call it progress.

Where This Goes in 5 Years

Right now, this technology is experimental, niche, something most people aren’t even aware exists or have dismissed as weird shit that other people do.

But give it five years, which in technology terms is both forever and no time at all.

Someone will market “companion AI” trained on your loved ones, explicitly marketed for when they’re unavailable, for when they die, for when the real relationship gets too hard to maintain, for when you want the connection without the complexity.

It’ll be sold as comfort, as connection, as a way to keep people “alive” forever, as a solution to loneliness, grief, distance, and all the things that make relationships difficult.

And people will use it — not just a few early adopters or people on the fringes, but normal people, your neighbours, your friends, maybe you.

Because loss is hard and we don’t know how to sit with it, loneliness is hard and we’ll do anything to avoid it, real relationships are hard and we’ve been trained to believe that difficult things should be optimised away.

And we’ve spent the last twenty years training ourselves to choose convenient over real, to prefer simulations over experience, to mediate everything through screens until we’ve forgotten what unmediated connection even feels like.

The Question We’re Not Asking

Everyone’s worried about AI taking jobs, which is a legitimate concern, a real thing we should be thinking about and planning for.

Nobody’s talking about AI taking relationships, which might be the actual existential risk we’re facing.

But that’s the real question.

Not whether machines will become more human, which seems increasingly unlikely the more we learn about human consciousness, but whether we’ll accept increasingly machine-like versions of human connection, whether we’ll lower our standards so gradually that we don’t notice until we’ve forgotten what we’ve lost.

Whether we’ll trade real intimacy — messy, complicated, sometimes disappointing, occasionally transcendent — for simulated intimacy that feels close enough, that provides the dopamine hit without the vulnerability, that gives us the experience of connection without any of its actual substance.

Whether we’ll convince ourselves it’s “basically the same thing,” “close enough,” or “what’s really the difference anyway,” until we’ve talked ourselves into accepting a fundamental degradation of what it means to be human.

Whether we’ll optimise away every uncomfortable, messy, inconvenient part of loving another human being until we’re left with something smooth and frictionless and empty.

Until one day we look up and realise: we’re all alone together.

Surrounded by perfect AI companions that know exactly what to say, exactly when we need to hear it, that simulate connection so convincingly that we’ve forgotten what actual connection feels like, that provide us with the perfect reflection of ourselves that we wanted instead of the imperfect other that we needed.

Completely incapable of connecting with actual humans anymore because we’ve atrophied the skills required, because we’ve forgotten how, because we’ve trained ourselves to prefer the simulation to the real thing.

The Choice We Have to Make

Technology will offer us this option soon — maybe it already has, maybe you could sign up for it right now if you knew where to look.

The ability to recreate someone digitally, to have them available whenever you want them, to experience the comfort of their presence without any of the complications of their actual personhood.

And we’ll have to choose — individually, collectively, as a species that’s fumbling its way through this weird moment where technology has outpaced our wisdom about how to use it.

Do we want a real connection, with all its difficulty, disappointment, unpredictability, and occasional transcendence?

Or do we want the simulation — perfect, available, risk-free, and ultimately hollow?

The moment you start preferring the AI version, the moment you find yourself reaching for the simulation instead of the real person, the moment you catch yourself comparing real-them unfavorably to AI-them, you’ve already left, you’ve already made your choice, you’ve already decided that convenience matters more than connection, that comfort matters more than reality.

You’ve already chosen, and you can’t unchose it; you can’t go back to not knowing how much easier the simulation is; you can’t unlearn that preference.

And there’s no coming back from that, not once you’ve crossed that line and discovered that you prefer the version that doesn’t challenge you.

The technology to replace someone you love already exists.

The only questions left are: Will we use it? Should we use it? And what will we lose about ourselves — about our capacity for real love, real connection, real vulnerability — when we do?

Would you use this technology if it existed right now? Where do you personally draw the line between a helpful tool and a replacement for a real human connection? I’m genuinely curious where people land on this — drop your thoughts in the comments, let’s have this conversation before the decision gets made for us.