Comment by knotimpressed
1 day ago
How much of this was written with/edited by an LLM? Some parts of it feel somewhat Claude-ish, and I’m very curious if that’s a false positive.
Super cool work! I had no idea UWB tags had gotten so cheap, or that CR2077 was a size.
I don't think CR2077 is a size, it may be a typo for CR2477 which is very common?
CR2477 is high enough capacity that it's used in a number of zigbee or zwave800 based products that have no wired interface.
https://www.google.com/search?client=firefox-b-d&q=cr2477+ba...
https://aichiplink.com/blog/CR-Battery-Sizes-and-Specificati...
It is indeed a typo! Thanks.
Interestingly I’ve used Zigbee for well over a decade, tried easily over a hundred different devices, built others homes, and I’d never encountered this battery size until this project!
Will fix the typo later
Over half, apparently. https://www.salahadawi.com/hacker-news-ai-detector/49947472
Well that’s fascinating. The answer is a small amount was Claude written, like one of the image captions when I had it find the render and insert it, and I had it do a brief grammar and spelling check, but very little. Most of those chunks that tool has flagged were 100% handwritten.
FWIW, we've been finding that tools like grammar checkers far more imprints on text than people intend when they use them (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...). It is one of the two most common reasons why text gets classified as genai even when the author didn't use tools to generate their text. (The other is automatic translators.)
s/claude/codex