221–230 of 313 posts
Earlier quoted context omitted.
If you’re judging my writing so shallowly, I don’t think I’m writing for you.
> If you’re judging my writing so shallowly, I don’t think I’m writing for you. No, you are writing for people who see LLM-signals and read on anyway. Not sure that that's a win for you.
I really like Oxide's take on AI for prose: https://rfd.shared.oxide.computer/rfd/0576 and how it breaks the "social contract" where usually it takes more effort to write than to read, and so you have a sense that it's worth it to read. So I get the frustration that "ai;dr" captures. On the other hand, I've also seen human writing incorrectly labeled AI. I wrote (using AI!) https://seeitwritten.com as a bit of an exp…
My biggest sorrow right now is the fact that my beloved emdash is a major signal for AI generated content. I've been using it for decades now but these days, I almost always pause for a second.
Earlier quoted context omitted.
I just skimmed the article, but I can already tell it's chock-full of LLMisms. In other words: ai;dr Edit: ok, I've checked your profile and now I see that this is your website that you're astroturfing every thread you reply to. Stop doing that.
I’m seriously not astroturfing, point is serious. Most people here would not be able to tell the difference between good human and good AI generated writing anymore in a blind test, especially a blend of both. So if the reader often can’t tell, why is the source of a well written, interesting piece of writing more important, than the effect of the content itself? I don’t feel the OP made a good argument imho.
You say that, but I think it's really, really obvious in this case.
If you disagree, you should show us both versions.
Earlier quoted context omitted.
AI-generated code is meant for the machine, or for the author/prompter. AI-generated text is typically meant for other people. I think that makes a meaningful difference.
> Programs must be written for people to read, and only incidentally for machines to execute. from the preface of SICP.
Earlier quoted context omitted.
What is creative about generating an article from a stub? The kernel of the article around which the LLM constructs the content? I'm not trying to be an ass, just curious.
The stub itself. So why not just read the stub? As my other post said there is more value in an article than just the creative idea.
When a human writes it, yes.
Earlier quoted context omitted.
I rarely see LLMs generate code that is less readable than the rest of the codebase it's been created for. I've seen humans who are short on time or economic incentive produce some truly unreadable code. Of more concern to me is that when it's unleashed on the ephemera of coding (Jira tickets, bug reports, update logs) it generates so much noise you need another AI to summarize it for you.
The main coding agent failure modes I've seen: - Proliferation of utils/helpers when there are already ones defined in the codebase. Particularly a problem for larger codebases - Tests with bad mocks and bail-outs due to missing things in the agent's runtime environment ("I see that X isn't available, let me just stub around that...") - Overly defensive off-happy-path handling, returning null or the semantic "empty"…
To add to this list:
- Duplicate functions when you've asked for a slight change of functionality (eg. write_to_database and write_to_database_with_cache), never actually updating all the calls to the old function so you have a split codebase.
- On a similar vein, the backup code path of "else: do a stupid static default" instead of erroring, which would be much more helpful for debugging.
- Strong desires to follow architecture choices it was trained on, regardless of instruction. It might have been trained on some presumably high quality, large and enterprise-y codebases, but I'm just trying to write a short little throwaway program which doesn't need the complexity. KISS seems anathema to coding agents.
I really like Oxide's take on AI for prose: https://rfd.shared.oxide.computer/rfd/0576 and how it breaks the "social contract" where usually it takes more effort to write than to read, and so you have a sense that it's worth it to read. So I get the frustration that "ai;dr" captures. On the other hand, I've also seen human writing incorrectly labeled AI. I wrote (using AI!) https://seeitwritten.com as a bit of an exp…
Years ago I wrote something similar to test a biometric security piece that used keystroke timings (dwell and stroke) to determine if the person typing the password is the same person who owns the account. Short version of a long story is that it would be trivial to get data for AI to reproduce human typing. Because I did it years ago using something only slightly more sophisticated than urandom .
"Growing up, typos and grammatical errors were a negative signal. Funnily enough, that’s completely flipped for me." For me too and for writing it has the upside that it's sooo relaxing to just type away and not worry about the small errors much anymore.
I laugh every time somebody qualifies their anti-AI comments with "Actually I really like AI, I use it for everything else". The problem is bad, but the cause of the problem (and especially paying for the cause of the problem)? That's good!
It's not as one-dimensional as good vs bad. Transformers generally are extremely useful. Do I want to read your transformer generated writing? Fuck no. Is code generation/understanding/natural language interfaces to a computer good? I'd have to argue yes, certainly. I cry every time somebody tries to frame it one dimensionally.
Earlier quoted context omitted.
My biggest sorrow right now is the fact that my beloved emdash is a major signal for AI generated content. I've been using it for decades now but these days, I almost always pause for a second.
We're in the brief window of time when AI's writing style is the weirdness. It's an artifact of the production process, like JPG blur, MP3 distortion, autotune's rigidity. And it didn't take long for those things to become normalized, in fact for them to become artifacts that people proudly adopted and embraced. DJs release tracks built from MP3s samples instead of waves. Autotune is famously a 'sound' that was once…
Also, you forgot the extremely enervating: "It's not X. It's Y. ."