I don’t mean this in an accusatory way, but did you write this with AI? I feel I’m becoming paranoid and seeing AI authorship in everything. The structure matches ChatGPT output.
Clankers were trained on the writing style of individuals such as myself (ASD). While I do consciously (and subconsciously) code switch, I’m aware I have a default “sounds like ChatGPT” voice when dealing with technical discussions, especially if I’m trying to be precise or guard myself from accusation or attack.
I’m ok with it, but I’m now going to autism at you / over explain it, partially because I think it might help you parse the difference between human and machine when reading these things.
You asked in apparently good faith and you deserve a full explanation.
The underlying pattern comes from a perversion of the “measure twice, cut once” mentality; I create the crux of the argument, forecast likely objections, rewrite to close off said objections, sand the edges off, check if I was unintentionally offensive, check if I presented the facts to the best of my ability, check for logical fallacies, then finally sweep to see if there is any ambiguity or obvious attack surfaces left. Then I read it out loud to myself.
That mode flattens everything into a “safe, palatable, high signal to noise ratio, use dot points so people don’t lose you, don’t write like yourself” style.
(And I still miss typos sometimes. That actually really, really irks me).
Anyhow, when I said the clankers copy us, I didn’t mean just vocabulary. Expand the CoT (chain of thought) the next time you use ChatGPT; you’ll see they made it do this exact same process.
PS: You’re not the first to point it out either. It’s one of the reasons I dubbed my blog Clanker Adjacent
I don’t mean this in an accusatory way, but did you write this with AI? I feel I’m becoming paranoid and seeing AI authorship in everything. The structure matches ChatGPT output.
No.
Clankers were trained on the writing style of individuals such as myself (ASD). While I do consciously (and subconsciously) code switch, I’m aware I have a default “sounds like ChatGPT” voice when dealing with technical discussions, especially if I’m trying to be precise or guard myself from accusation or attack.
I’m ok with it, but I’m now going to autism at you / over explain it, partially because I think it might help you parse the difference between human and machine when reading these things.
You asked in apparently good faith and you deserve a full explanation.
The underlying pattern comes from a perversion of the “measure twice, cut once” mentality; I create the crux of the argument, forecast likely objections, rewrite to close off said objections, sand the edges off, check if I was unintentionally offensive, check if I presented the facts to the best of my ability, check for logical fallacies, then finally sweep to see if there is any ambiguity or obvious attack surfaces left. Then I read it out loud to myself.
That mode flattens everything into a “safe, palatable, high signal to noise ratio, use dot points so people don’t lose you, don’t write like yourself” style.
(And I still miss typos sometimes. That actually really, really irks me).
Anyhow, when I said the clankers copy us, I didn’t mean just vocabulary. Expand the CoT (chain of thought) the next time you use ChatGPT; you’ll see they made it do this exact same process.
PS: You’re not the first to point it out either. It’s one of the reasons I dubbed my blog Clanker Adjacent