New post: "On not becoming a cyborg" https://nolanlawson.com/2026/08/31/on-not-becoming-a-cyborg/
Some thoughts on how LLMs are influencing me, and how I'm trying to fight back against that influence.
6
Replies
0
Boosts
@nolan Appreciate that your still wrasslin’ with all of this.
@davatron5000 "Wrasslin'" is a great word 😆
@nolan Thankful my local accent is mostly model-proof.
@nolan Nice piece. Part of the difficulty in this is that those word and style choices are there because they were already popular, so like..load bearing is definitely one you can probably find me saying too. So then do you counter it by abandoning those? Its tricky.
@bkardell Thanks! I'm not sure it's true though that Claude is just giving the "average" of everything it's read. It seems to be tuned to pack a lot of meaning into very dense sentences. Some agents' "thinking" or inter-agent communication also seems to be a weird kind of pidgin optimized for token efficiency (see e.g. https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/#~1200-agents-sent-%3E70,000-messages-and-files-on-an-unsanctioned-message-board,-and-~700-attacked-hugging-face). So I think it has its own unique voice.
@nolan Yeah I'm definitely not saying it is giving the average either, I'm just saying that the ones you mentioned are, afaik, all reasonably popular somewhere in tech writing at least, and that's probably where it comes from. So for me, I almost struggle with whether I need to change how I write to avoid sounding like the thing that actually sounds like me in terms of word choices sometimes.
