Reading time: 6.14

1,482 words

Hey Reader,

One technology, two futures — and how we’re letting the wrong one win…

Imagine putting your head into a box which has six speakers in it. From each speaker comes a different noise: several conversations, music, cutlery clinking, traffic sounds, laughter, footsteps, a dog barking, a child crying. All at full volume.

Someone taps you on the shoulder and you turn to listen. You can see their mouth moving but they could be saying rhubarb rhubarb rhubarb for all you know.

You nod and smile and learn later that they were telling you about the death of a loved one, so then YOU want to die, too, because you must have seemed incredibly insensitive.

That’s what it’s like for me in a crowded room. You can concentrate and focus on the conversation in front of you, no matter how noisy it is. I know you can, because I watched you do it last week at a busy event, while I panicked quietly on the edges.

That’s also what it’s like for people who use hearing aids, like my mum. Hearing aids amplify everything indiscriminately, as does my brain.

AI as a lifeline

So imagine now how exciting it is to hear about an AI development that can mitigate this. There’s research that explains it in detail but the gist is: they monitored people’s brains as they paid attention to different conversations and sounds, then programmed an AI to read those brainwaves in noisy situations, helping the person to focus on the sound they wanted.

In a 2023 study at Columbia University, researchers found that AI-powered hearing aids using EEG signals reduced background noise interference by up to 40% for users in crowded environments.

This would literally transform my life. It would enable me to take part in business events. I’d be able to come across as who I am rather than the way I do at the moment, which is absolutely not me at my best. I’d find social situations much less stressful and exhausting and anxiety inducing.

AI’s potential to solve real, tangible problems — like auditory overload and food distribution — makes its misuse even more infuriating because while some people are using it to heal, others are using it to steal.

AI as theft

NOW imagine my reaction when I saw Samantha Harman’s post about how someone fed her brilliant book Just Get Dressed into an LLM (large language model) — because let’s not call it an AI; it’s not intelligence, it’s an aggregator — in order to pitch her PR services.

Take a beat and consider this: the person couldn’t be bothered to read Sam’s book, so they fed it into an LLM and asked for a summary. Then they got the LLM to write the pitch email. Then they sent it. And presumably patted themselves on the back for a job well done.

I don’t know how to say this any more plainly than:

THIS IS NOT OKAY.

Not just because Sam’s book is now being used to train LLMs without her permission and without any recompense. And not because it’s going to “replace” authors like her (it won’t).

Not all LLMs operate this way. Some, like Hugging Face’s “Open” models, use opt-in datasets where creators explicitly consent to their work being included. But that’s not the case here. You can read all about the class action lawsuits being brought against Meta and other LLM companies here. I wish them luck and godspeed and hope they take the big corporations to the cleaners. Again, not because LLM-produced books will squeeze real authors out of the market; I don’t believe that will happen. Because these LLMs are stealing our work and creativity then selling it back to us without us seeing a penny of the profits. Which is late-stage capitalism all over, really. Exploit, extract, profit — but profit for a tiny few.

As of May 2026, over 12 publishers have joined the lawsuit, with estimated damages exceeding $2 billion.

AI as a crutch

Anyway, those things do concern me, but my main concern is for your brain and your identity.

Often, we can tell when someone’s used an LLM to produce something — or at least, I can, after years of editing. But the real issue is the hollow result. LLM output lacks the spark of lived experience, the quirks that make writing yours.

Don’t misunderstand what I’m saying here: I don’t for a second believe every “AI tell” is actually AI. The problem is that LLMs have been trained on US, so of course they’re going to use constructions we use. It’s the combination of those tells with sentence structures, vagueness, and surface-level “thought leadership” that gives it away. I mean, you can pry my em-dashes from my cold dead hands. I was using them before the internet was a sticky stain on a tech-bro’s sock.

My biggest problem is the way we are not just losing our voices but actively giving them away. And even worse: for the sake of “ease” we are trading our ability to think.

And if you would accuse me of ableism for railing against this, I need you to listen to what I’m actually saying because — as I opened this article with — AI could be a real force for good if we stop wasting our time with LLMs.

Yes, LLMs lower the barrier to entry for writing, but so does a voice recorder. The difference is that a voice recorder doesn’t pretend to think for you, nor does it regurgitate other people’s work as your own.

It’s ableist to suggest that it’s “too difficult” for autistic or adhd people (like me) to do the hard work. Or that dyslexics can’t write because they’re dyslexic. Bullshit.

I’m not saying neurodivergent and disabled people shouldn’t use tools. I’m saying we shouldn’t outsource our thinking — the part that makes our work uniquely ours.

AI as a tool, not a brain transplant

I use AI tools: Otter for transcribing meetings and pulling out the points I need. Transcription tools for taking my voice notes and turning them into prose I can edit. Spelling and grammar checkers I can check my work with, and make sure I haven’t inadvertently plagiarised anyone. I use it to help me edit: “Show me where there are gaps in my logic and research, where I’ve repeated myself, where the structure makes no sense.”

But I don’t prompt it to start my work. To come up with ideas. To do my thinking for me. To take my bullet points and write a piece from those bullet points because what comes out isn’t even close to me, no matter how much I “train” it on my voice. I know, because I’ve tried it.

Then when I realised how much more difficult it was becoming to think, I stopped and ran in the other direction.

The bargain is not what you think it is.

What is it costing us when we can’t be bothered to even skim read a couple of chapters, take in some of the book reviews, and think up our own approach to an author? When we think the best thing to do is simply regurgitate inaccurate quotes and stuff them into a half-arsed pitch?

What’s it costing us when we say, “Ugh writing is hard” then outsource the thinking part to an LLM that will shrink every single original thought you ever had into a McDonalds beef-adjacent patty?

Writing IS hard.

Thinking IS hard.

IT’S SUPPOSED TO BE.

Because out of the hardness comes something brilliant and unique and innovative. Eventually. After we’ve done the damn work. After we’ve spouted a bunch of crappy boring surface level words and thoughts, we start getting to the interesting ones. We start figuring out what we REALLY think, then articulating it in a way that only we can.

Those words we make up? The turns of phrases that sound a bit wonky but are instantly recognisable as us? The Vickyisms and Youisms? That’s what makes our work interesting, and memorable, and it’s what people will seek out in the growing filthy ocean of AI slop.

So by all means, they can keep plugging bland prompts into an LLM and slurping it back out again. Telling themselves that it’s hard for them in particular and carrying on.

Leave me behind.

PLEASE, for the love of all that is holy, leave me behind. We’ll be over here being as human as we can possibly be, and we’ll be the ones who get ahead because of that.

TTFN,

Vicky 🫡

p.s. Know someone who might enjoy this email? Please forward it to them and get them to sign up here.