Oh, whoops, I appreciate you catching that. It’s fixed now.
Thanks. ^^
Originally that heading was going to be Person-Generated Overview, to more directly contrast with “AI”-Generated Overview, but I thought that version still sounded too… mechanical. So I wound up settling on “handcrafted” instead.
This topic has been on my mind a LOT lately. For my professional work, I use LLMs all-day, every day. It’s nearly required by my employer. And for the work I do, the speed at which I can build features and solve problems is impossible to ignore.
I am torn by it though. I spent almost 30 years building my skillset in this software engineer world. I don’t want to lose that ability. But, there’s an aspect where it might make more sense for someone like me to use an LLM rather than someone who is learning these skills for the first time. I have the privilege of over half my life building the judgment, discernment, and taste to make, hopefully, smart choices about what the LLM generates. That’s a decidedly different situation than someone sitting down at a computer for the first time and reading an HTML/CSS tutorial instead of just telling the clanker what to do. That’s how I started and there is immeasurable value in that learning.
On the other hand, I think many of us are excited by the prospect of more people being able to express themselves in all the ways we have found value in. Art projects, enthusiast zines, shoutings into the void, all those ways the small web adds to the world. Gatekeeping that is something I wouldn’t want to engage in without more thought.
A final thought that might not land: I am a devout Catholic and the Church has come out rather strongly against the use of AI in creative pursuits. Yet, they recognize that Pandora’s box is open. We have to find a path that allows us to be discerning in how we use these tools. That’s hard work and it will look different for each of us. I’ve been thinking along these lines around the ideas of “givens” and “artifacts”. Imagine a person painting a landscape. The landscape is a “given”. That is, it is given via the act of Creation. The painting is an “artifact” derived from the “given”. It’s the result of human endeavor. LLMs, and some other activities, move us farther away from the “givens” since they create “artifacts” from other “artifacts” ad infinitum. As our work moves farther away from God’s creation, we lose the soul of what makes it meaningful.
I don’t know where, or even if, I’ll land at some definitive stance. It seems that in this current stage of AI, the firmament shifts under our feet each week.
I don’t want to come off as confrontational, so I’m sorry if I do, but I’m honestly getting very sick of hearing that AI, especially Gen AI is breaking some sort of ‘gatekeeping’ in creative spaces. It’s really not, it’s actually making it very hard for artists and other creatives who are just starting out, be they ‘professional’ or not.
There is nothing creative about skipping the actual creative process, in fact every time I hear about this supposed gatekeeping that artists do, it usually comes from people who have no actual interest in the creation of art, the analysis of art, or the value of art (I do not mean you, I’m talking more broadly about people engaged in the more technical parts of the web that often see art simply as content). Art requires one or more of the following to actually be art, in my mind: skill (more accurately style, I don’t find skill that important), intent, and emotion (the emotion can range from it moves you to tears to it just looks cool). Gen AI is incapable of all three (even style, it steals form actual artists, often without consent). But guess who is? Anybody who tries anything creative, even if they lack skill and are a total beginner. So, no, there is no gatekeeping and it’s absurd to claim that a machine built on stealing and devaluing art is a solution to it.
When it comes to programming I am totally inexperienced so I cannot say whether the use of LLMs is appropriate or not, I’ll trust your expertise on that. However, I’m sure if your hard work was taken without your consent and fed into an automaton that is made to replace you (whether it actually produces better work or not), you’d probably find some issue there.
I’m an atheist myself but I appreciate your look at AI creations through the Catholic lens. I also find that the further we move away from things that are very human, what has soul (especially in the arts), the worse it is for us mentally and emotionally. Again, I apologize if this response is a little heated but it’s something I’ve been really worried by.
No offense taken at all. I agree with you. While I do believe that coding is a creative pursuit, I believe that our intentions are important. Maybe a better word for coding is “craft” rather than “art”. But, perhaps, that all rest on the intentions. When I’m asked to build a feature at work, it is 100% a commercial endeavor. Build the feature, sell it, make money. Art is rarely, if ever, a part of the conversation. But, when I write some code that allows me to express myself on a web site, that feels closer to art. Same tools. Same programming language. Decidedly different goals.
That might be an artificial line that I’ve created to justify my own activities 
As for something closer to what we would normally think of as art (digital or not) such as drawing, painting, design, sculpting, I agree with you that AI doesn’t allow for anything resembling creativity or art. It simply lacks access to the “givens” (to build on my earlier point).
The gatekeeping I was referring to, a bit imprecisely perhaps, was more around the technical aspects of hosting a website or getting domains hooked up. Those things are part of being creative in a web space but can sometimes be tangential to the expression a person is trying to share. In the worst case, it’s enough of a barrier to stop them all together. Those are the gates that I might be more comfortable with AI helping to open.
This is a really interesting perspective!
I’m not religious at all, but some of my feelings about AI “content” end up sounding like I am. Sometimes I want to yell vague feelings that amount to, “it doesn’t have a soul!” But I don’t believe in souls in general, so that argument falls apart when coming from me specifically. 
The idea of AI-generated “content” being an artifact of an artifact does make sense to me, though. Based on how LLMs work, I don’t think that AI stuff has actual meaning, just the appearance of it. But the whole point of writing and art is to communicate something. Even a baby crying because they’re hungry conveys more meaning than an LLM text.
I’m a strong proponent of “If no one could be bothered to write it I don’t see why I should spend my limited time on this earth reading it.” LLMs are a curiosity that should never have left university labs, the sheer amount of resources needed to make them sort of useful, the intellectual theft, the deskilling of users. It’s Embrace, Extend, Extinguish for software developers and if the corps have their way, for thinking.
This is one of the arguments I was addressing in my previous post, which I can’t tell if you’ve read.
I actually really like your approach to coding as craft, shame that often times it’s relegated to being pure function and nothing else. I can see it being a creative outlet just as well.
I see what you mean about gatekeeping, it’s not exactly what I thought you were referring to. I can see how it would be easier to make your way through the web with help from AI but I still think that by nature LLMs/Gen AI consume and chop up the actual work of others and using them is still more detrimental to a beginner than helpful. In any pursuit, knowing the foundation is the most important thing, in my opinion. Even if you don’t become advanced in a field, having that base is essential and the only way to get there is to not skip the learning process.
This is exactly my position too and oh how I wish we really left a bunch of stuff in uni labs…
I hadn’t and thank you for reposting it. I really like the point you’ve made. It’s definitely why I put the word “might” in italics. I’m not convinced of my own point. Your argument makes a lot of sense to me and I love the framing you’ve used of a guide or a mentor or a teacher adapting their feedback to the needs of the person asking the question.
There’s certainly a lot of the “why” missing when a chatbot does something for you. Sure, a user can ask it why and it will dutifully explain for the most part. But, that explanation isn’t tailored to the person or the situation which is the point I took from your post. I agree with you that’s quite valuable.
Well. As someone who isn’t a dev and doesn’t work in anything remotely related to coding but loves the indie web, maybe my perspective in this is way too skewed, but… I do use Claude.ai to help me with css/html stuff for my blog/website. I usually screw up something (why isn’t this fading effect workingggg) and the AI comes in handy in showing me the issue. It’s usually something stupid, I’m a newbie after all. All my previous knowledge of css/html came from looking at other ppl’s codes in geocities and tumblr themes 20 years ago. 
But when I’m using Claude, I do have to make a point of instructing it to explain it to me like I’m a 12 year-old kid who doesn’t know any better, because otherwise I have no idea what it did on the code, it would just give it to me corrected but without a lot of commentary… or with commentary I absolutely do not understand bc I have no idea what I’m doing most of the time. lol
I also read somewhere that no piece of code is “original”, it’s all about “stealing” and “recycling” and sewing together bits and pieces from different sources, which I don’t know if it is true. But if it is, it’s very different from how people from the arts work. Usually you can tell when an artist takes inspiration from another artist (and sometimes it can cross the line of plagiarism if you’re not careful), but I don’t see a lot of talk about “plagiarism” when it comes to coding.
All of that to say that yes, AI does help me bridge a knowledge gap I have to do things on my website and blog. Sometimes I even ask it to explain something that I still didn’t understand even when reading it from W3school. (I’m a slowpoke when it comes to learning code, folks… but I like making things pretty. 
)
I know for a fact there are coders out there who consider AI code plagiarized, and for some applications that may very well be true – I don’t know enough to judge. But I feel like there’s a large gap between “asking the robot to check my HTML” and “asking the robot to build an entire program from scratch, which I will then monetize for a quick buck”.
I can accept people trying to write their own code and then seeking AI assistance in debugging when something goes wrong. They’re still trying to do their own work and using the LLM as assistive technology.
I would exercise caution using AI on a personal site. The same way LLM-generated text has certain “tells” it likes to overuse, the code (particularly css) is the same way. You’ll notice if you look at different vibe coded websites they all have a very common look. You don’t want that ending up replacing the personalness of your own site.
I agree, but I think there’s a difference between having a LLM generate code for you and writing code yourself, testing it, and passing in your code with a prompt explaining what you want to accomplish and asking it to explain why the existing code doesn’t work.
Thank you.
That can be part of it, though that wasn’t the main idea I had in mind. The point about guides (as in mentors) is more specifically about guides missing out on skills and experiences – and how missing out can be detrimental and discouraging.
The language is different, and some expectations are different, but there’s still some conceptual overlap. Code can be copyrighted and proprietary. Talk of copyright and licensing isn’t usually as relevant on the indie web, though, where many people give out tutorials and templates for free, so here’s how I like to think about it:
What can I do to honor the people who have helped me?
For the time being I’ve gone about this a few different ways. On my personal site, I’ve linked to general resources on the About page, the Web Neighbors page, and the Useful Links page. I may not be writing a bibliography for every single span and div, but I’m putting enough out there that if a newbie visits my site and is like “hey, how can I make something like that?” then it’ll be easy enough for them to discover those links.
Besides introductory coding guides, I also want to credit the folks who’ve helped me create things that aren’t quite so elementary and straightforward as making a link or a div, and here I’ll give a more specific example to illustrate.
When I first started my personal site, I knew I wanted a filter system for my essays page, but initially all I could find was this tutorial that relies on Javascript. Since I knew that some people browse the web without Javascript, I kept looking for a CSS-based solution. Eventually I found this CSS-based filtering tutorial, which I used to overhaul my Essays page, and later I would stumble across Solaria’s version, which I used (and credited) as the basis for a template called Bibliodex. Since then, other people like J have used Bibliodex and credited me for it (and even recommended it to others). So now, through that chain of credit links, a person browsing any one of those other sites could follow the link trail back to @solaria. I value the creation of these link trails both as a way to help others and as a way to recognize the ways that other people have helped me.
Other people are also helping you when you get code out of a LLM, and the difference is that the LLM creates a layer of obfuscation. You don’t get to know who is helping. You don’t get to know whose data in the dataset was used to generate the output, and you don’t get to know who to credit for it. This is the kind of thing I was talking about in the section of my post on surfability.
I still worry that the output is likely to be inaccessible, even if beginners are just using LLMs to debug their code rather than generate it outright. There’s so much misinformation about accessibility out there on the internet already, and any LLM trained on that will just perpetuate it.
Hi! Thanks for replying. After I commented here, I read your post (it’s already on my postroll too), and I totally see and agree with your point.
As a newbie myself, I absolutely seek out credit pages on personal websites I visit. There’s often stuff I see on their sites that get me all “how did they do thisss!!” and I’m crushed when there are no credits or even mentioning how they did what they did (which is pretty common in websites created by experienced devs but I can’t hold that against anyone lol).
I compulsively collect these resources links on my bookmarks dump site. (I’m already including the stuff you linked on your post there for later reference btw.) As a newbie, I think it’s super important to share with other newbies how to do this and that on the web. I even created a little page on my website talking about customizing status.cafe bc when I created my own account I couldn’t find anything about how ppl made their pages cute.
I got pretty upset reading the surfability part of your post exactly because that thought that had already crossed my mind – if I don’t post my question somewhere searchable (reddit, stackexchange, wherever), are other people going to miss out in finding the same answers later on? But I kind of dismissed it thinking that my issues/questions were waaaay too basic to be considered relevant, that the solutions were already shared ad infinitum across the web and there was nothing special about it.
I did, however, kept thinking that even if that were true, I myself would be unable to know if one specific issue that I have with my code would be interesting to share because I literally do not have the ability to assess that.
Having said that, your post made me go back to these qualms I had. I might just turn to random subreddits from now on though (stackexchange is not newbie-friendly lol), or simplify the code when I’m stuck and comment on it later on my blog. It’s very easy to just use the LLM, but it’s really one of those “the easy way isn’t the right way” situations.
Thank you, Folkmoss. I appreciate you hearing me out – that’s not something I take for granted.
For what it’s worth, when I wrote that post, I was also thinking about how beginner questions deserve to be recognized as valuable unto themselves. Just earlier this week, for instance, I had an interaction with someone who was asking for help with designing a page layout. We talked through some troubleshooting with CSS grid, centering a div, and adding scrollbars, and I introduced them to working offline with code editors. That conversation is one that I’m glad I got to have, and as a novice myself, I enjoyed getting to pass along some of the things I’d learned.
I hope you get to have that experience too someday, if you haven’t already. You might find that it can be really rewarding.
By the way this forum does have a help channel, feel free to post any code problems you have there! Beginner questions are definitely welcome! (infact I would be able to help more if they were beginner haha)
Your posts here also made me make sure I still had w3schools linked on my homepage (which I still do), as that was such a big help for me when I was starting to learn HTML and CSS. Their little code editors embedded in almost every page helped me so much by letting me see examples and tinker with them.