Gary Lucero Writer

Category: Communication

  • The Architect of My Own Reality

    The Architect of My Own Reality

    I’ve spent the day vibe coding an app that has allowed me to create something that mirrors my unbridled enthusiasm and creativity. I know that sounds stupid, but I realized there’s no one around to tell me this isn’t possible.

    And since I’ve yet to reach my $20 a month limit in Cursor, and the app works as I’ve designed it to work, I’m happy. So, I decided to write about why vibe-coding is so much better than my day job. But instead of writing it myself, I let Gemini 3.1 Flash Lite do it for me, based on my prompt, of course.

    I also edited it a bit:

    In my day job, I am a professional software tester. My life is defined by the friction of the “No.”

    I spend my hours documenting elegant solutions, identifying critical edge cases, and proposing features that would make the user experience seamless. I send these tickets off into the void, only to have them return with a status update that feels like a cold splash of water:

    Won’t Fix. Out of Scope. Resource Constrained.

    In the real world, software is a compromise. It’s a series of negotiations between what is ideal and what is budget friendly. I am the person who sees the potential, but I’m not the person who holds the purse strings.

    Then, I close my work laptop and open my personal terminal. I start to “vibe code.”

    The Infinite Canvas

    When I vibe code, the dynamic shifts entirely. The hierarchy of the corporate office dissolves. There is no product manager to tell me that my vision is “not on the roadmap.” There is no lead developer to sigh and tell me that the architecture won’t support the change.

    When I am vibe coding, I am the business analyst, the coder, and the user all at once.

    If I decide that a button should glow when hovered, it glows. If I decide that the data should flow through a custom-built pipeline that defies standard conventions, I build it. And if I hit a wall—if the laws of physics or the limitations of a specific library make a feature technically impossible, I deal with it.

    The Absence of “No”

    The beauty of vibe coding isn’t that everything is easy; it’s that the intent is never rejected.

    When I encounter a limitation, the programmer, the version of myself that exists in the flow state, never says “No.” Instead, the conversation becomes a creative pivot.

    • The Real World: “We can’t implement that feature because the legacy database doesn’t support real-time updates.” (The conversation ends.)
    • Vibe Coding: “The database doesn’t support real-time updates? Okay, let’s build a middleware cache or switch to a different engine. Let’s find a way to make it happen.” (The conversation evolves.)

    In this space, a limitation is not a dead end; it is simply a prompt for a different path. I am never told that my idea is invalid. I am never told that my curiosity is a waste of time. I am never told that the project is “good enough” as it is.

    The Joy of Unrestricted Creation

    There is a profound, almost meditative joy in this autonomy. Vibe coding allows me to reclaim the agency I lose during the nine-to-five. It turns the act of programming from a chore of compliance into an act of pure expression.

    When I am vibe coding, I am not just testing the work of others; I am manifesting my own. I am learning that while the world may have its constraints, my imagination does not. And as long as I am the one at the keyboard, the answer to “Can we do this?” will always be the same:

    Let’s see how we can make it work.

  • AI Can Be Useful

    AI Can Be Useful

    Artificial intelligence is a much hated technology and for good reason. Like all the advances before it, it’s being leveraged by big tech as the be-all, end-all solution for all the world’s problems. It can lower costs, drive revenue, prevent fraud, and do it all cheaply. 

    Or so we’re led to believe.

    To be honest, though, AI isn’t bad. It’s the people who use it poorly and worsen our lives with it, they are the problem.Fortunately, we too can use it, and we can use it wisely.

    There are hundreds of uses for AI, but here are a few I find it particularly good at.

    Building Software

    Proficient coders will tell you AI is a superpower, but only if you know what you’re doing. I agree that’s the case when creating production software, but I create software for myself. 

    In my case, I vibe-code the entire project, and I never look at the source code. I ask Cursor Agents, my IDE of choice, to write the code and tests, investigate issues, etc., and I serve as the business analyst and software tester. 

    It does the rest.

    I also create and configure the GitHub repo, but Cursor Agents handles creating branches, committing the code, and pushing it. 

    And it does an excellent job. None of my projects are huge, and all the coding is done with either the Auto model or using Grok 4.5 High, but it very inexpensively serves the role as code creator and maintainer. 

    Recommending

    I recently chatted with Gemini about books. I told it the kinds of books I like to read, giving it several examples of ones I have enjoyed, as well as some basic information about what I like and dislike. I then uploaded a list of all the Kindle books I hope to one day buy, and asked Gemini to review the list.

    It pointed out how many of them didn’t meet the criteria I had outlined, because they were too technical, or written poorly, or there was some other issue with them. And it recommended several other books. I let it know which I’d already read and liked, added a few it suggested to my list, and now I have a pinned conversation in Gemini I can use anytime I need a book recommendation.

    This is an activity that AI is custom-made for. You could, for example, give it a list of TV shows or movies or video games that you like, and several you don’t, and ask for suggestions. And if you’re already familiar with ones it suggests, go back and forth with it until it understands your preferences and suddenly you have a valuable recommendation machine.

    Sorting Thoughts

    There are days when I feel upset, or have questions on my mind, or maybe I’m just wondering how something in the world works. In these cases I often turn to AI. I tell it what I’m thinking, what my problem is, or what my question is. We chat, exchanging comments, suggestions, and questions until I’m satisfied.

    This is another one of those situations AI is good at. No, it can’t solve all my problems, but it’s excellent as a sounding board, and it has a vast array of knowledge. I do this often, most of the time using the interactive feature in Gemini, where it’s always on, listening for my input, then it responds.

    We chat until I’m satisfied.

    Teaching Me

    I always have lots of questions. 

    I’m constantly wondering about technology, products, gardening, and more. I often turn to AI in these situations, providing it with pictures or other information. It almost always provides solid answers that help me understand the world we live in.

    I’ve gone from knowing almost nothing about how to set up a repo in GitHub properly to actually understanding the difference between CodeQL and Dependabots. I now understand better how to care for the plants in my yard. I have a better understanding about how smart phones, versus bridge cameras or full-frame cameras, work.

    AI encompasses the world’s knowledge, and if you are careful enough, and aren’t too naive, it can help you understand it.

    Reminding Me

    I couldn’t survive without the reminders I keep on my Android phone, or the calendar events I have for appointments. They keep me from forgetting all of the various things I need to do.

    But sometimes I just need a quick way to remember something. With Gemini, I can tell it things, and later, ask it to remind me about a fact I committed to it. I can do the same with notes in Google Keep, or in Google Drive, and Gemini can find those too. 

    It’s a persistent memory store that I can turn to at any point on my phone or computer and retrieve information I would otherwise lose to my aging mind.

    Assisting with Video Games

    Video games are an important aspect of my life, and have been for over 40 years. But my middle-aged brain, eyes, and hands don’t work like they used to. 

    Enter AI.

    Not only can I ask for game recommendations, but it also helps with understanding how accessible and/or difficult a game is, how long a game is, where I’m currently at in the game, and occasionally, it helps me get through a game. 

    In those cases though, my personal opinion is, if I need a walkthrough to complete a game, it’s not worth playing.

    Since chatbots like Gemini have access to all of the FAQs, reviews, opinion pieces, walkthroughs, and other information on video games, it’s the perfect place to answer all my questions.

    And So Much More

    This of course isn’t all AI can do. 

    I use it constantly, and while there are times it’s frustrating, where maybe I say, “Hey, Google,” to my smart watch or smartphone and it doesn’t respond, or I ask it a question and it returns something that doesn’t resemble an accurate response.

    But when it works, it’s like magic. It provides access to so much information, and does it so well so much of the time, it has become indispensable.

  • Why I’ve Embraced AI

    Why I’ve Embraced AI

    I know the implications of AI. The fact that running it wastes enormous amounts of electricity and water. That it could potentially take millions of jobs. And it could one day destroy the world if we decide to let it run it and it decides to, say, nuke us all.

    In the here and now, AI helps me with cognitive load. I can discuss issues of all sorts with it, from technology to social security, music, video games, movies, TV, art, and anything else that comes to mind, and it helps me make sense of the world.

    It’s not my friend. I’m not stupid enough to believe that it has some sort of consciousness, and I often stop short of answering its questions because it doesn’t matter to it if I respond or not. 

    But it still helps.

    I figure I will use it while I can. It could be in a year or so the big tech companies realize they have to make a profit and pull a Disney Plus on us, raise the prices substantially, and I no longer have access. But for now, I will continue to use it.

  • The Most Annoying Thing About AI is How You Can’t Convince It of Anything

    The Most Annoying Thing About AI is How You Can’t Convince It of Anything

    There are certain individuals who are so self-centered, so self-involved, they assume they know what you are, who you are, what you think, and why you think it. You try to convince them otherwise, but you always fail.

    That’s a chatbot in a nutshell. You converse with it, wanting feedback similar to that of a reasonable human being, but it takes what you say at face value, and regardless of how much effort you put into correcting its view, it never changes.

    Why converse with a chatbot in the first place, you say? That’s a good question.

    There are some topics the human beings I know have no interest in, so, if I want to talk to someone about those things, I turn to Gemini. I’ve had many satisfying conversations with it, so it’s not like it’s completely unable to fill in for real humans.

    But it often falls short.

    I know when to cut off a conversation with Gemini. It asks me a question, and I know it doesn’t actually care about my response. I’m not so stupid I think there’s a sentient being on the other side of the screen.

    But sometimes, I just want it to understand what I’m saying. Really understand. Help me to understand myself. But it can’t bridge those gaps, look me in my eyes, see that I’m pleading with it to get me, and then make a leap of faith.

    It just can’t do it.