[dmje]

This writing had an interesting effect on me. I went in knowing nothing about the app or author and frankly having no need for whatever it was he might be selling. But by the time I finished reading I was significantly amused and interested that I’m going to go check his stuff out.

What I’m trying to say is that this is someone who can really, really write - he’s deeply funny and self deprecating, but obviously also knows his shit, big-time. And that’s a massively powerful skill, maybe as much of a skill as being able to write Swift or make great interfaces or ship an app.

> “If you grew up with Tamagotchis, you already understand why this was tempting. Not the “cute pixel pet” part. The part where a device the size of a digestive biscuit turns into a low-resolution hostage negotiator.”

This is irritatingly good and it makes me want to buy his products and subscribe to his RSS feed. Great writing is powerful magic.

[anonymous908213]

Funny, that was around the point in the article where I was beginning to get irritated reading it because it felt like reading LLM output. LLMs love melodramatic headers ("THE CHILDHOOD TRAUMA"), outlandish and not particularly coherent metaphors ("hostage negotiator"), the overly terse arrow constructions that I've never seen a human write in my life ("something that feels less like “open app → consume lesson” and more like “tap creature → it looks at you → you do a small thing together”"), the segue into a redundant list of bullet points, the pointless not x but y ("The blob wasn’t a mascot here, it was the interface") which poorly establishes a contrast where it doesn't make sense to.

The funniest part to me is that I suspect the LLM generated the line about the 4th of July, and the suspected prompter being British, felt the need to insert an explanation for why "they" would reference it, in a voice/cadence that doesn't really match the rest of the article:

> "Confetti, fireworks, the whole 4th of July experience (I've seen it only in movies though, not sure why but it's not celebrated in the UK)"

I can't definitively say this is LLM-generated, but it resembled it enough so that I still came away annoyed for having read it.

[integralid]

I guess to each their own. I enjoyed the style and even laughed a bit at the part you highlighted (writer humorously pointed out the obvious fact that 4th of July is not something UK celebrates).

I think you look for AI too hard. Perhaps that kind of dry humour is not too your liking, or you're not used to this style? FWIW i lived in the UK a bit, so I'm rather familiar with the way locals speak casually.

Btw. you can check his pre-chatgpt writing style, for example [1]. Looks similar enough to me!

[1] https://drobinin.com/posts/things-i-learnt-in-2021/

[anonymous908213]

To be clear, I found the dry quip about the 4th of July amusing, and specifically pointed out that I thought that specific parenthesised line was inserted by the author. I don't think a British author would naturally reach for "4th of July" as their frame of reference for bombastic celebrations in the first place, though. My point was that seemed to be something the LLM generated and the author riffed off of.

I'm not about to go into a deep dive analysing the author's past writing style, but there is a clear difference just from glancing at the headers alone. Looking at older articles, such as this "featured" one[1], they all share a commonality: the headers are boring. Matter-of-fact. Plainly descriptive. "The reasoning". "The background". "The research".

[1] https://drobinin.com/posts/what-ive-learnt-after-sending-147...

Then a sudden spate of activity in late 2025 after years of not having written anything other than yearly recaps, and all of the new posts share a different commonality: the headers are 'creative'. "The Childhood Trauma". "Teaching a circle to care". "47 seconds: a villain origin story"[2]. "The uncomfortable engineering truth".

[2] https://drobinin.com/posts/how-i-accidentally-became-puregym...

It is quite a noticeable shift to go from always writing useful headers that clearly communicate the purpose of the following text, to always writing clickbait headers that try to hook the reader's emotional attention.

[integralid]

Fair. I understand where you're coming from and you have some good observations. Investigating this in depth is probably not a good use of our time, but who knows, maybe the text is indeed AI assisted? In this case kudos to you for being having a sharp eye and being vigilant. This thought didn't even cross my head.

[dmje]

Wow, that's interesting - maybe I'm totally naive about spotting AI generated writing but my gut feel was it is incredibly human - didn't even ponder for a second that the robots had touched it! Will have to go re-read...

[jstanley]

Not only is it pretty obvious why US independence day is not celebrated in the UK (although maybe that was tongue-in-cheek?), but we do have a fireworks night on a different date.

[WindyMiller]

The interface equivalent of ten sad fireworks and a pack of sparklers in a rainy back garden probably wouldn't have the same addictive effects, though.

[nrhrjrjrjtntbt]

Yes I would want this person designing my app because it is clear they are very curious and into the craft.

[qweiopqweiop]

The OP has replied it was mostly AI written... I guess we're well past the em-dash era now.