3 ms·

the docs were clearly not written by a human, so proceeding from the expectation of global coherence, or that of adherence to principles that would be implied t

the docs were clearly not written by a human, so proceeding from the expectation of global coherence, or that of adherence to principles that would be implied to a human being reading this in a certain context, is likely to end up being a waste of time

90s_dev 1y ago
Wait, was this language guided by AI???
- fwip 1y ago
  From my personal judgment: definitely using AI for the code & commits, probably for the documentation, no idea about the design. For instance, no human would write this commit message, with the preformatted-code-block: https://github.com/helixlang/helix-lang/commit/4d949efd42b8dce47bc02578c6ff5c54d4144241 https://github.com/helixlang/helix-lang/commit/4d949efd42b8d...
- ze7111 1y ago
  Hey I'm on of the devs; No, the code is not by AI: the current compiler and , corelib's code, the thought process and language design is all human, AI would produce slop in that case, however, English isn't our first language, so things like README's, Docs and such, had the following process, we wrote an initial draft, with broken English not cohesive at all - but, the content is factual, then sent it to ChatGPT to improve it, did it use buzzwords and such - yes, but at the time, with our level of English it did seem really good, looking back now it does seem really bad, but Commit messages are done using GitLens Commit Message Generator - simply because I was too lazy to write each and every commit message, Moving forward, We have already started to rewrite the README and the Docs, by hand, we will first commit that, then do a revision with AI, while maintaining the tone and formality, only improving grammar and punctuation. AI is a really good tool when it comes to improving English and communication, however I do stand with the fact code should NOT be written by AI.
lostdog 1y ago
"Comprehensive Error Handling: Helix's error handling system is novel to helix, it allows for either complete error handling or a more pythonic approach to error handling, this allows for more control over how errors are handled in Helix." Wow, these really do seem to be LLM written, and I would guess ChatGPT. This long README has a bunch of generic, redundant statements over and over again.

90s_dev 1y ago

Wait, was this language guided by AI???

fwip 1y ago
From my personal judgment: definitely using AI for the code & commits, probably for the documentation, no idea about the design. For instance, no human would write this commit message, with the preformatted-code-block: https://github.com/helixlang/helix-lang/commit/4d949efd42b8dce47bc02578c6ff5c54d4144241 https://github.com/helixlang/helix-lang/commit/4d949efd42b8d...
ze7111 1y ago
Hey I'm on of the devs; No, the code is not by AI: the current compiler and , corelib's code, the thought process and language design is all human, AI would produce slop in that case, however, English isn't our first language, so things like README's, Docs and such, had the following process, we wrote an initial draft, with broken English not cohesive at all - but, the content is factual, then sent it to ChatGPT to improve it, did it use buzzwords and such - yes, but at the time, with our level of English it did seem really good, looking back now it does seem really bad, but Commit messages are done using GitLens Commit Message Generator - simply because I was too lazy to write each and every commit message, Moving forward, We have already started to rewrite the README and the Docs, by hand, we will first commit that, then do a revision with AI, while maintaining the tone and formality, only improving grammar and punctuation. AI is a really good tool when it comes to improving English and communication, however I do stand with the fact code should NOT be written by AI.

lostdog 1y ago

"Comprehensive Error Handling: Helix's error handling system is novel to helix, it allows for either complete error handling or a more pythonic approach to error handling, this allows for more control over how errors are handled in Helix." Wow, these really do seem to be LLM written, and I would guess ChatGPT. This long README has a bunch of generic, redundant statements over and over again.