[moWerk]

Fair suspicion in 2026 — but we actually started drafting that announcement in early 2023 before first committing it here: https://github.com/jrtberlin/aos2.0-post. As a non native speaker i sure ran my commits through grammar polishing multiple times over the years. And i am genuinely curious now what to avoid to not sound like an LLM if you could dig out one or two examples.

[retsibsi]

I wonder if they meant this HN post (the OP), rather than https://asteroidos.org/news/2-0-release/? The latter doesn't seem LLM-written to me, but the HN post does.

[antonvs]

Correct.

[moWerk]

Full disclosure, i spent ~3 hours crafting this post to hit the tone i wanted to convey and am kind of proud of the wrist-size linux banger i came up with. I am usually not good with writing since it takes me ages and i might be a bit over sensitive right now. All i wanted was to spare you all the rocky grammar as a cherry on top. To now find that the polished version triggers your "its (completely?) Ai written" sensor. Lesson learned i guess.

[bjackman]

FWIW as someone who sadly has about 10% of his brain constantly engaged in AI detection when I read HN, absolutely nothing about the announcement struck me as feeling AI generated.

(I feel guilty for pointing it out but I guess it's actually a compliment in this context: I even saw a beautifully reassuring mis-conjugated verb elsewhere on the website. I wonder if LLMs will have to start injecting these errors to give us an authentic feel. Maybe they already have).

I think we are at a point now where just coz AI is so prevalent, every post on any programming forum will have at least one comment saying "AI slop".

[antonvs]

Read the text of the HN post. If it’s not AI slop, then it’s a canonical example of the training data that produces the slop.

[Aachen]

I hadn't opened the article yet and was just browsing comments over my cereal when I saw this and thought "ugh, amother one?" and went to check for myself

I didn't get an LLM vibe at all. Looking for it specifically, the bullet point about UI improvements is a candidate; the sentence following "mediawiki" could be an autocompletion; maybe the first sentence of the download section... but they're also all plausibly just a bit 'marketing team worded', so not necessarily LLM-sounding. And even if an LLM made suggestions to some small parts like these, who cares? There aren't any slop sections that waste your time, this is just like using a thesaurus — if these parts were LLM suggestions in the first place, which I don't actually expect because then there should be more of it

This type of post very poorly lends itself for auto-writing anyway. It'll put emphasis on the wrong aspects and not come out as intended, at least in my experience it's more work coaxing it to good results. It can be helpful not to start from a blank page but that's about it, I rarely find a sentence among the output that's fully usable as-is

[antonvs]

I was referring to the text of the HN post. Go read it.

[Aachen]

Aye-aye o7 ...

I saw that one already and didn't suspect slop there either

[antonvs]

I was referring to the text of the HN post. The second and third paragraphs in particular, but most of it, really. If you didn’t use an LLM for that then congratulations, you’ve successfully nailed the Silicon Valley LinkedIn marketing guru zeitgeist which LLMs seem to gravitate towards.