This is very bad. No filtering, all transmissions on harmonics.
Emotes in readme, emotes in scripts, emotes in source files, lol libs used = LLMed up the ass. Someone pointed Claude (yes, CLAUDE.md is in the repo) at https://github.com/F5OEO/rpitx and told it to wrap it around in fluff go code, then sprinkled profanities and pirate theme on top.
I don't recall many, if any, Github repos containing this emoji-vomit before the rise of AI, and likewise natural human conversations in forums and such were also not like this, so I find it very odd and distinctly unnatural. Where did this "vibe coded" style actually originate from?
I occasionally saw a readme with a couple of Emoji in it, but the behavior where every title, every bullet point, every sentence ends with one or more of them is utterly obnoxious. There's something about the sources that chatGPT uses which causes it just sprinkle Emoji absolutely everywhere, it's one of the most obvious tell tales that it came from chatGPT without a prompt to tell it otherwise.
It's especially odd, as the publicly available conversations between real humans that I know are being used for AI training don't contain anywhere near as much emoji. Twitter/X might come close, but that content also comes with plenty of imperfect grammar and spelling. I wonder if ChatGPT was specifically prompted to this style by default.
It's just very obviously made with claude, from the style, to the commits of tens of thousands of lines of code a day, to the parts where Claude commits something sensible and the author goes back to add in curse words. Nobody has ever developed software in the same way that claude tends to, where suddenly a whole readme appears in a commit fully formatted and filled with emojis.
I agree with you it's Claude, but I do commits like this. I develop and iterate in the private repo, then update huge (logical) chunks in the public repo.
I share some code/solutions, but I don't have to share the entire process, minute details, etc. So one day there's a 5 kB MVP python script doing something, and 3 months later it turns into 15 kB module with pretty classes, wrappers, docs etc.
This is very bad. No filtering, all transmissions on harmonics.
Emotes in readme, emotes in scripts, emotes in source files, lol libs used = LLMed up the ass. Someone pointed Claude (yes, CLAUDE.md is in the repo) at https://github.com/F5OEO/rpitx and told it to wrap it around in fluff go code, then sprinkled profanities and pirate theme on top.
I don't recall many, if any, Github repos containing this emoji-vomit before the rise of AI, and likewise natural human conversations in forums and such were also not like this, so I find it very odd and distinctly unnatural. Where did this "vibe coded" style actually originate from?
I occasionally saw a readme with a couple of Emoji in it, but the behavior where every title, every bullet point, every sentence ends with one or more of them is utterly obnoxious. There's something about the sources that chatGPT uses which causes it just sprinkle Emoji absolutely everywhere, it's one of the most obvious tell tales that it came from chatGPT without a prompt to tell it otherwise.
It's especially odd, as the publicly available conversations between real humans that I know are being used for AI training don't contain anywhere near as much emoji. Twitter/X might come close, but that content also comes with plenty of imperfect grammar and spelling. I wonder if ChatGPT was specifically prompted to this style by default.
I don’t see a CLAUDE.md (or any mention of Claude) in the repo that’s posted or the one you quoted?
https://raw.githubusercontent.com/psyb0t/piraterf/88c5fc416d...
That's because it's in the .gitignore.
It's just very obviously made with claude, from the style, to the commits of tens of thousands of lines of code a day, to the parts where Claude commits something sensible and the author goes back to add in curse words. Nobody has ever developed software in the same way that claude tends to, where suddenly a whole readme appears in a commit fully formatted and filled with emojis.
I agree with you it's Claude, but I do commits like this. I develop and iterate in the private repo, then update huge (logical) chunks in the public repo.
I share some code/solutions, but I don't have to share the entire process, minute details, etc. So one day there's a 5 kB MVP python script doing something, and 3 months later it turns into 15 kB module with pretty classes, wrappers, docs etc.
It's like Usenet but with emojis.
I miss those days, zero snowflakes, all trolls. Now we have more verbiage in CoCs than code.
This is just a slopped version of https://github.com/F5OEO/rpitx
This stuff shouldn’t be on HN.
How are you doing output RF filtering? Fixed filters or some kind of adjustable one?
There is no filtering. Apparently if you just say
> Built for engineers who understand that good RF practices matter more than arbitrary administrative boundaries
then actually following good RF practices is optional.
Not at all: https://github.com/F5OEO/rpitx/blob/master/doc/Tunerpitx.png
why not PSK31, JT-8 etc?