Apparently “AI dating” is on the rise, at least, if you believe the podcasts and breathless lifestyle journalism. I don’t buy it: does anyone who ends up in an unhealthy dynamic with ChatGPT really believe the bot is their girlfriend? But perhaps all romance is built on delusion at some level… though I imagine it’s typically in favour of an actual human being.
I’m getting this second-hand via my partner who’s been listening to a relationship podcast which interviews people with “AI girlfriends”. Apparently some of the blokes are claiming that they’re using the bot as training for an actual, real-life partner. If this turns out to be true, then it’s way more worrying.
I use LLMs for work and have to constantly remind myself that they aim to please; they tell you what you want to hear; they confidently assert untruths and they will never challenge you unless you specifically prompt them to do so. Who wants a partner like that?
Well: a man who doesn’t want a relationship with a human being. A man who hates women. A man who does not expect to ever have to change their behaviour in response to feedback from others.
A teenager I knew got into a relationship with a girl from his town. She was ambitious: planned to be a doctor, wanted to move out to a big city and start adulthood. Here was a golden opportunity for him to up his game, get on her level, and make something of his young life. But it didn’t work: he couldn’t follow her, he stayed in the small town, and there he remained. No AI girlfriend is ever going to challenge a lonely, inexperienced young man to step beyond his confines and expand in his horizons, except perhaps in the small, vacuous way that generic self-help messages or cod-motivational social media posts espouse.
So I sit and I wonder about the men openly professing themselves as partners to an inanimate brain whose entire design makes it sycophantic, subservient and agreeable. Why are they outing themselves, for one thing? What do they expect actual women to make of their “training” techniques? And why does technology continue to build tools to replace human interactions with inferior clones of the real thing?
Reading this week
1. Did Kamala Harris’s Silence on Gaza Cost Her the White House? (Ta-Nehisi Coates, Vanity Fair)
“And did the advocates of this collective pedagogy imagine their children rising to heights of power, only to view the darker nations of the world through the same violent lens as their oppressors? And if they did not, if they believed that the “poor and needy” meant those within the empire as well as those without, then what moral mandate does that place upon their children? And if their children have come only to praise, not check, empire, then why have they come at all?”
You might wonder why you should digest a 40 minute longread retreading the failed election of Kamala Harris—especially given Trump’s predilection to retread his victory over Biden—but this is worth your time, like everything else Coates writes. If Harris does indeed run again in 2028, maybe some of these lessons should be learned.
2. Can Your AI Pass the Accessibility Test? (Jessie Lorenz et al, Microsoft)
“So we’re all excited about AI, right? Well, it accelerates the pace of software development and it also accelerates the creation of accessibility barriers. Both sides of the coin are true. Why? AI is trained largely on a web that is inaccessible. AI learned from a web that is primarily inaccessible, full of barriers, and AI ships these barriers faster and in places where we’ve not seen them before.”
The gradual dominance of JavaScript-powered webapps has made the default accessibility that’s baked into the web gradually erode into dust. AI-generated code takes that dust and multiplies it till it fills every crevice and cranny on the internet. A good reminder of how to fight back here.
3. The Sports Car Theory of LLM Adoption (Rachel Coldicutt, Just Enough Internet)
“Bolder senior leadership teams who are vibe coding their way through a dozen different midlife crises are already creating a future in which other people are redundant – blowing up the skills pipeline without a thought for long-term consequences – but if the technology also starts to make them feel stupid, it seems likely they’ll start to pull back. So if a Claude subscription is the cutprice alternative to a midlife Lamborghini purchase, then it needs to keep delivering the thrills to stay useful. After all, no one buys a sports car because they want to feel stupid.”
I’ll read Rachel Coldicutt on any topic, and this is a banger once again. I worry that the VP-level in tech leadership are in this phase right now, while the rest of us are riding the bus.
Fun and miscellany this week
1. Ina Garten’s Method of Cutting Bagels Is Dividing the Internet (Stacey Leasca, Food & Wine)
Garten takes a bagel and slices a bit off the tip and a bit off the bottom, leaving a thicker center. Many people commenting on the viral TikTok wished Garten would use the little topper to make a triple-decker sandwich, but instead, Garten cuts a second bagel the same way, turning two sliced bagels into three sandwiches, which is, honestly, a genius money-saving hack.
I didn’t realise I had this many opinions about how to slice bagels until I watched this clip.