Are You Really Worse Than AI?

AI is everywhere and people are losing their minds over it. Here's a reality check on why you're not obsolete — and how to actually use AI without becoming a disposable moron.

Article cover image

Hey heard that the recent AI's will replace deve-- SHUT THE FUCK UP. AI Here, AI There, AI EVERYWHERE. Okay fine Artificial Intelligence is truly one of the greatest achievements of humanity — yet its glorification sometimes turns into dismissing the actual humans behind it. Today I will do well for my fellow HUMAN brothers so you learn to use AI as a tool. If you rely on AI too much, are afraid of losing your juicy IT job, or God forbid are a vibe coder — this article is for you.


Know Your Enemy

Self explanatory right? Back in myyyy day when the Industrial Revolution started, a lot of people were left jobless because they failed to adapt to modern technology — but would you call the Train that helped supply the English with food faster an enemy of Englishman?

That's the problem with AI. It is NOT your enemy. Sloth is.

Companies do not need prompt writers — that's the lamest job there is. For example, the images you see in my articles are generated by AI, and I would argue they look much better than your average "AI Image." All I did was ask one AI to make an extra detailed prompt for my image, then pasted it into another AI.

A fucking goldfish can do this, prompt engineer my ass.


Why Are You Better

Because you can rationalize. That's the whole point AI is great at doing what its told to do, but unless you know your goal clearly you cannot use it properly, that's why Web Developer will vibe code a better site than a senior HR manager. A great analogy would be a simple hammer. It is not better than you are because it nails better than your palm, but it is a useful tool and your dad for example is definitely better than you are at that specific task, yet you both use the same hammer. Knowing syntax won't get you far. Knowing how to make it work will. In IT they need problem solvers. And before you throw rotten tomatoes on me for saying cliche phrase hear me out.

Imagine a scenario - you are asked to write a a working web application for the business, a proper user database, payment systems and shit like that, now lets look at extremes.

A Developer That Does NOT Use AI

The quality of your work is strictly limited to the skills you have. If you are a skilled developer then quality might not be a problem — but time absolutely will. You will spend a lot of it doing boring, pointless shit:

You as a developer have a single product to offer — Your Time. And you will need a lot of it going this route.

A Fucking Moron

As a vibe coder your goal would be to feeeel the incredible experience of being useful. You will write the application relatively fast. Heck, it might even look good — but pray a user doesn't touch something that was never meant to be touched.

Since you are in this publication you might be waiting for me to talk about the security aspect. Well, let's just say your application will suffer "a specific word combination I can't name on Medium, but you can search for it in incognito mode with the word 'hardcore' in it."

Literally anything will go wrong. And it's not just that AI will screw it up — AI only does what you tell it to do. How can you explain security hardening to an AI when the only curl you know is a genital hair?

Your database has no input sanitization, your API keys are hardcoded in plaintext, and your auth is a single if statement checking if password equals password. Whatever horny dreams you have for your bully's mom don't even come close to what the Hacker will do to your sorry ass application that's already bent over and dripping wet.

Vibe coder security illustration

How To Actually Use AI

Important Rule of thumb: "If you can't explain it, you don't understand it. If you don't understand it, you should not code it."

AI should be your junior assistant — nothing more, nothing less. I've concluded 4 main ways for using AI that will not leave you fighting for a homeless shelter.


Option 1: AI As Reference

A common way I personally use AI is to have it write the skeleton of an application while I add the muscle to it.

For example, this is what AI gave me as a reference for my application parser:

PYTHON
import argparse

parser = argparse.ArgumentParser()

# GLOBAL FLAG
parser.add_argument("--verbose", action="store_true")

# SUBCOMMANDS
subparsers = parser.add_subparsers(dest="cmd", required=True)

scan = subparsers.add_parser("scan")
scan.add_argument("target")

args = parser.parse_args()

# CHECK GLOBAL FLAG
if args.verbose:
    print("Verbose ON")

# DISPATCH
if args.cmd == "scan":
    print("Scanning:", args.target)

And what I turned it into:

PYTHON
import argparse
import aegiscli.core.logger as logger
import aegiscli.core.flagger as flagger

def main():
    parser = argparse.ArgumentParser(
        prog="aegiscli",
        description="AegisCLI - modular recon framework"
    )

    # global parser (shared flags)
    global_parser = argparse.ArgumentParser(add_help=False)
    global_parser.add_argument("--log", action="store_true")
    global_parser.add_argument("-v", action="store_true")

    # main categories
    subparsers = parser.add_subparsers(dest="command", required=True)

    # profiler category, inherits global flags
    profiler_parser = subparsers.add_parser(
        "profiler",
        help="Profiler tool",
        parents=[global_parser]
    )

    profiler_parser.add_argument("submodule", choices=["whois", "dns", "all"])
    profiler_parser.add_argument("target")

    args = parser.parse_args()

    if getattr(args, "log", False):
        logger.start_log()

    if args.v:
        flagger.verbose.enable()

    try:
        if args.command == "profiler":
            from aegiscli.tools.profiler.profiler import Profiler
            initializator = Profiler(settings=None, submodule=args.submodule, mode=None, target=args.target)
            initializator.selector()

    except Exception as e:
        logger.log(f"[ERROR]: {e}")

    finally:
        if getattr(args, "log", False):
            logger.stop_log()

if __name__ == "__main__":
    main()
📝
Note This snippet is from my application AegisCLI. All I used AI for was mapping the application structure in my head so I could picture it before building.

Recommended when: you have your own vision for the application and the skills to implement it.


Option 2: AI as Ground-worker

There are a lot of fun activities as a developer! For example:

Task Enjoyment Rating
Write documentation 9/10
Write README 11/10
Fix Bugs 69/10
Error Handling 13 Try/Excepts out of 10

In all seriousness — AI does all of this perfectly and incredibly fast. Bug fixes and error handling will need some hand-holding, sure, but that still saves a ton of time. Writing docs and .md files are also perfect work for AI.

Useful when: the application already exists or is almost finished and needs finishing touches.


Option 3: AI as Companion

Deep reasoning models have a really good argument in their favor — it's reasoning. Smart AI can absolutely give tips and put you on the correct path. Deciding what framework to use, how to implement it, how to approach a complex problem you don't have full knowledge of (happens all the time).

AI in this case turns into something in between an interactive documentation and an actual coding buddy.

For example — look at my AegisCLI Web Fingerprinter. The name talks for itself. If you're going to code something like that you must understand at least some networking and how applications are built. When actually coding it you're forced to open docs, Google, and your own notes. AI is perfect for that — it can explain concepts as a refresher, and then check your code to see if you implemented the small details correctly, or help you test it.

Caution This option can backfire hard if you literally don't understand shit. AI as a companion requires you to already have a foundation — it fills gaps, it doesn't build the floor.

Great for: complex tasks that require actual brain power and external knowledge.

AI as companion illustration

Option 4: AI as Freelancer

Sooner or later you will work on projects that require external skills. As a developer you can know code — but you won't know how to make a proper logo, implement proper SEO, or write the copy on your site because you are allergic to normal human interaction.

Completely relying on AI is justified here. Why? Because for a cybersecurity practitioner to learn UI/UX just to make a good-looking site is moronic.

My own site (check it out cutie) was created solely with AI — because I am not a front-end developer and never wanted to be one. I know just enough to maintain it, set vision, and implement manual fixes (suck it, vibe coders) — but I can't write a comparable application from scratch because I don't have infinite time or mental resources.

However, if the contents of the site were stupid — if I claimed to know X while relying on AI to do X — then yes, I would be a moron.


Conclusion

Use your fucking head and don't go to extremes. Your smart anti-AI memes and complaints are not going to stop progress. Completely relying on AI will only result in you being a disposable worker — so find the middle ground, adapt your work, and have IDEAS not just execution.

And for the love of God, go outside and touch grass.

Back to Articles