Just ONE more AI think-piece, I know I know…

Artificial Intelligence (AI) has {insert your preferred platitude about its rise}. There are think pieces, hot takes, LinkedIn manifestos, panic, hype, and the occasional thoughtful post—every emotion’s on the table. And yeah, the tech space is fully saturated with AI content. Well, here is another entry. These are just my thoughts on the phenomenon, and how I, as a freelance web developer, (spoiler) use it.

With all the noise around AI, here is what this article is NOT going to be:

In this piece, I will give my thoughts about MY experience with AI, how I use it, and my overall philosophy—where I give my take on common discussion points around it.

AI as a Learning/Planning Tool

The area that AI has helped me with more than anything is general learning. I love a good tutorial as much as the next person but sometimes having the extra (virtual) hand for support is helpful, especially when I am trying to get down to the true “why” behind what I am learning. Sometimes in development it isn’t the actual code portion of a project that will be cumbersome, but the things on the margins, like setting up a compiler and properly enqueuing files that end up taking up more time without assistance. That being said, here are a few prompts I like to ask AI when I am trying to jumpstart my learning.

Things I like to ask AI when I am learning

“Take me through this code line-by-line and explain what every part is doing”

There are times where I do rely on AI for code solutions or minor tweaks to filters/hooks that I do know exist. However, rather than blindly copy/paste the code, I will usually have AI take me through the code and explain what each part is doing before fully integrating into my work. AI doesn’t always understand the full context of your work, so this is a good way to not only learn but to avoid unintended gotchas. An addendum to the prompt I sometimes like adding “explain this to me like I’m 5” for an even more thorough explanation.

“If I didn’t have AI how would I come to this solution”

Similar to the previous question, I like to not only understand the solution, but also where the solution comes from. There is an old adage in web development: “The difference between a senior and junior developer is the senior developer knows how to Google better”. While that is a little simplified, the essence is true. Most of the time what comes with experience is not knowing more code, but being able to FIND the right code. For this reason, I like seeing the steps that I would have to take to get to the solution. This usually leads me to the correct documentation that I can then reference later for similar features and builds.

Using AI for Brainstorming

Whenever I’m tackling a non-conventional build—something beyond the typical static website—I definitely lean on AI during the planning phase. In web development, there are usually multiple ways to build the same thing, each with its own pros and cons. When trying to figure out the best approach, I find AI really helpful for organizing my thoughts and making decisions. I like to weigh factors like security, performance, and scalability—that is, how well the solution will hold up over time. While I can research all these on my own, AI lets me throw out an endless stream of “what if this?” and “what if that?” scenarios. It doesn’t replace the final call, but it does help clear up the inputs and considerations.

For example, I was debating whether to keep using Bootstrap as my main CSS framework or switch to something like Tailwind. I’d seen plenty of think-pieces and LinkedIn takes on both sides. One of the main sticking points was code weight—Bootstrap has a lot of dependencies and extra stuff, while Tailwind is more minimalist. After going back and forth with AI on the details, I realized I could stick with Bootstrap, which I’m more comfortable with, and slim it down by only including the parts I needed. In fact, after trimming, Bootstrap actually ended up weighing less than Tailwind.

This is a good example of using AI to dig deeper into the “why.” I could have kept reading articles and weighing opinions, but having AI help made the decision process way more streamlined.

Code Improvements

Outside of learning, AI is certainly helpful in building out projects and coding. There are varying degrees to which AI is used from minor assistance to building fully built out applications. In this section, I don’t endorse or condone any sort of AI utilization practice (I get to that more directly later). Mainly, I want to touch on the areas that I have increased my use of AI in building my themes.

For Refactoring

Refactoring with AI is pretty common among developers. For me, I usually build out the initial feature mostly on my own. Once I confirm it’s working, I’ll turn to AI to help refactor or “clean it up.” This part can get tedious and prone to small clerical errors, so having a figurative second set of eyes is really helpful. Usually, I don’t just say “fix this {paste code}.” Instead, I’ll explain what the snippet does and what I want—for example, “I have this function that does X. I know it can be consolidated, and I need fallback checks in a few spots; can you refactor it?” Sometimes I ask for suggestions too.

I also like to drop code into AI and have it write detailed comments so I can easily understand it later. Honestly, I’ve started outsourcing a lot of the most tedious, error-prone tasks—typos, misspellings, formatting—to AI. That’s not to say I don’t get AI’s help writing code from scratch sometimes—more on that later.

ADA Improvements

A general overlooked part of code is ADA compliance. Some of it is missed by laziness (all coders have been there), time constraints, and most of all, lack of knowing every ADA rule under the sun. I have implemented findings from an ADA audit multiple times and have a pretty good grasp on it, and even I don’t know every attribute, rule, order, color, styling tags, etc to stay within WCAG guidelines.

In comes AI to save the day. I think AI does a great job of taking code that you have already written and enhancing it for screen readers. This includes small things like adding the appropriate “role” tag to a section to properly implementing a “focus:within” border from a styling perspective. And in the process, I have learned a lot more and am able to implement these changes AS I code rather than after the fact.

Less Reliant on WordPress Plugins

At baseline, I use plugins minimally and only if I absolutely have to. With AI, I am able to push the “if I absolutely have to” threshold further and further. Do I move away from plugins altogether? No. There will still be instances where I use plugins like Advanced Custom Field and Gravity Forms because re-creating their functionality is more trouble and less reliable (even with AI). And as long a plugin has a solid documentation repository, is regularly updated, and doesn’t add too much bloat, I am more than okay using it. 

Two recent examples where I would’ve normally used a plugin, but didn’t, were building an Activity Log and migrating a large batch of content.

For the data migration, I used AI to guide me through working with CSV files in PHP, something I hadn’t done before. From there, I used WordPress’s native functions to create posts, assign fields, and map relationships—no plugin needed. It ended up being faster, more controlled, and easier to tweak along the way.

As for the Activity Log, plenty of plugins can track user activity, but most either log too much, not enough, or make it difficult (or expensive) to customize. In fact, many of the advanced features are locked behind premium upgrades. With AI’s help, I built a lean, custom solution that logs only the actions I care about, in exactly the way I want, without the overhead or recurring cost of a plugin.

This is my favorite use of AI. Being able to dissect functions that plugins have and make them my own for the purposes of a project. And furthermore it allows me to say yes more confidently and sincerely when a client wants something that is truly tailored for their business – and in that, everybody wins!

Philosophizing about AI

There tend to be a lot of moral discussions around AI, especially when it comes to things like art, music, writing, etc. I can’t really properly speak to AI as a whole. The way I see it, it is just another workplace/corporate innovation and historically, those are always met with a certain level of fear, skepticism, radicalism, and divisiveness. And frankly, I am not qualified or skilled enough to make sweeping generalizations on what AI is or isn’t or whether it is good or bad. The parts I do feel comfortable having an opinion on is its impact on web development. 

Do you even have to know how to code? Can’t you just sit back and let it do it for you?

There is a lot of talk on the overall efficacy of code now and whether one even really has to know how to code to build products. Can AI just do it for you? And if so, do I even really need to pay for a developer? Looking at it the most reductive way possible, yes, AI can do it for you. And no you don’t need to pay for a developer. 

However, it will come with some drawbacks/limitations. First, AI doesn’t understand the granular context of your goals. This in some cases doesn’t matter. In other cases, context is definitely needed. Secondly, the technical output that AI creates is only as good as the prompter. Therefore, if you don’t know 100% what you are asking for, there will be a disconnect. Furthermore, if you don’t understand the code you are reading, the flaws may not be readily apparent (this is why I will usually have AI reference documentation and explain code to me). 

While AI technically can do things for you, the in-depth technical knowledge that comes from trial-and-error, annoying bugs, and the overall banging your head against a wall trying to figure out a feature can’t be replaced, and that ultimately is what will separate great developers from AI drivel.

Aren’t you afraid AI is going to take your job?

Many people have asked me whether AI is going to take my job. And by many people, I mean my parents and my own inner monologue. I think the fear is definitely valid. However, I have lost my job twice before AI became a thing, so… *shrug emoji*. This isn’t meant to be a cop-out, but more of a commentary that jobs/professions come and go, sadly. When web development started out in its infancy, it automated away plenty of jobs from non-technical people; so it is a special type of irony that web developers are now clutching their pearls that their livelihood is being put in danger by the same technological evolution that they benefited from.

While I don’t think that web developers as a whole will be eradicated from the Earth, I do think a certain subsection of them will. AND IT IS NOT JUNIORS—though the already uphill climb to get into software development does get steeper. And no, I don’t have any advice on that front. It stinks, and I don’t know how anyone gets a job in 2025. I am unbelievably lucky I switched into web development when I did in 2019, because that path I took can’t be recreated.

But I digress. In my opinion, to be a web developer today, professionals need to bring more to the table than superior technical knowledge. This includes, but is not limited to, the ability to manage projects, talk to clients, gather specs, make timelines, work with designers, possibly create designs yourself, etc. The days of just being a coding genius and that being all you need are over, especially if that genius comes with a lack of people skills. The ability to know every language, memorize every function, and crank out app after app is just not a competitive advantage as it used to be.

AI has made the barrier to learning code considerably lower; so if you are someone who bases their identity on technical prowess but is impossible to work with, people no longer have to accommodate that just because you’re smart or have a rare skillset. Ironically, as AI removes some of the human touch from coding, the value of human soft skills has only increased.

Personal Closing Thoughts

These next few statements are all true at the same time:

Zooming in on the third statement. This is one that kind of needles me. I hear all the time at developer meetups, “I use {insert your AI tool here}, and now I don’t even have to do anything.” And it is said with general excitement. I will be the last person to denigrate someone for taking shortcuts. However, not to be holier than thou, it does make me wonder why the people who say that started coding in the first place.

I switched careers into web development in 2019. And one of the main reasons is because I wanted to create and build things that were my own. I love the process of messing up, banging my head against a wall, and eventually coming out the other side. I love doing the menial tasks that eventually become vital to a big site; I love organizing my folders and giving the perfect names to my variables. And coding has provided a route to being a business owner, a very high level of fulfillment, and a creative outlet (even though I don’t know how to paint). 

At the end of the day, I am not here to cast aspersions on anyone and their use of AI. But in the process of benefiting from making our lives easier, I hope we as a web development community don’t lose the plot on why most of us started coding in the first place, which is to create, think, and solve problems. And as far as me losing my job—if the profession gets to a point where web development just becomes about AI prompting, and not about building and creating—they won’t have to worry about kicking me out. I will just step aside and tip my cap and thank coding for all that it has provided me.

TL;DR (written by AI)