We should promote this comments pattern. It means a human wrote it, not AI. AI often writes misleading or plain stupid comments, which is painful, especially if something is complex and important. When you notice this "human" tag, you know you can actually trust it more.

Aug 31, 2026 · 1:17 PM UTC

58
8
205
71,292
Sort replies: Relevant Recent Liked
Replying to @pie6k
What are the odds that if an LLM sees that in your codebase, it'll start adding it to its comments too?
3
63
5,278
You add an agent rule that if it sees that it is not allowed to modify it
2
7
4,168
more replies
Replying to @pie6k
Do I know I can trust it "more"? 😅
1
3
2,358
I feel very differently reading complex code I know was written by human. Somehow even for the best model i automatically suspect that if some code is really complex it might be full of slop and weird fixes or workarounds instead of proper architecture. Thus yes, I read something complex, well organized and made by human very differently, way more opened and way more optimistic i will learn something instead of just getting confused
1
6
1,992
Replying to @pie6k
"When you notice this 'human' tag, you know you can actually trust it more"
27
854
Replying to @pie6k
Wait you guys are inserting ai code, comments and all without reading it? At least clean up the comments come on man
3
210
Replying to @pie6k
A better solution: do not let LLM write any comments at all. If they are useless why are they in your codebase in the first place?
2
137
Replying to @pie6k
I do like the idea but this feels like it could be very easily abused
1
1
245
Replying to @pie6k
Wouldn't AI just learn to start tagging their comments with this after the AI training cycle comes around again?
1
194
Replying to @pie6k
it begs the question why let the llm make comments in the first place.
1
83
Replying to @pie6k
You trust humans to document the code correctly? I don't trust either.
142
Replying to @pie6k
I trust it more.
9
Replying to @pie6k
This gotta be bait
68
Replying to @pie6k
_what a depressing and shitty time to be a developer
255
Replying to @pie6k
If this goes viral I’m adding @human to my CLAUDE.md 😁
252
Replying to @pie6k
organic homegrown code
490
Replying to @pie6k
I"ve sent u a DM Adam
254
Replying to @pie6k
Good idea.. but won’t work for long
210
Replying to @pie6k
This is a good thought. But what is stopping AI Agents to use the "human" tag? There is no way to confirm if it's written by a real human.
164
Replying to @pie6k
AI will add `human` there :D
239
Replying to @pie6k
OMG i've been doing this!
166
Replying to @pie6k
I hope it’s not followed by an ai watermark
32
Replying to @pie6k
Post posted by #human ?
141
Replying to @pie6k
Better idea is to never write comments at all and don’t allow AI to write them either. Code should be self documenting.
1
148
Replying to @pie6k
I prefer NOTE(name):
4
Replying to @pie6k
Even better, just disallow the AI from writing any comments
68
Replying to @pie6k
137
Replying to @pie6k
ai will just write the @human tag
54
Replying to @pie6k
Why do I need a meat proxy to comment on code?
36
Replying to @pie6k
there have been at least dozen of clones of your app now, so many are open sourced now. how do you deal with it? are you doing anything different now that, there are so many in the town? i remember why you launched the app, have been following you since, then!
68
Replying to @pie6k
Can't we just have the AI add a human tag?
27
Replying to @pie6k
On the other hand, if I see that tag, I won't necessarily trust it
1
108
Replying to @pie6k
Jokes on you. I trust AI more than humans.
56
Load more