Site icon Day News

The New Job Of Being A Vibe Coding Cleanup Specialist Is Intriguing And Stirring Ample Controversy

The New Job Of Being A Vibe Coding Cleanup Specialist Is Intriguing And Stirring Ample Controversy


Know the ins and outs of becoming a vibe coding cleanup specialist.

getty

In today’s column, I examine the rising new job position of software professionals performing vibe coding cleanup. If you aren’t familiar with this unusual and rather controversial new practice, no worries, since I will be laying out the essentials for you.

Vibe coding refers to using prompts while in generative AI and large language models to have the AI generate programming code for you. It turns out that the AI-generated code often has various troubling problems, including being replete with errors, bugs, and sometimes evildoing hacks. Many assume that they can simply tell the AI to clean up any of the messes within the generated code. But AI isn’t fully adept as yet at doing such cleanup; thus, either the vibe coder or someone versed in dealing with programming code must laboriously inspect the code and attempt to make it correct and suitable for use.

A new role has been spawned. It is the rise of the vibe coding cleanup specialist. Will it last? Does it make sense? Who would take such a job?

Let’s talk about it.

This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).

Vibe Coding Is Hot

I will start by sharing with you the keystones underlying vibe coding. For additional detailed coverage, see my recent discussion at the link here.

The usual way to craft programming code entails the handwriting of source code, often using a conventional programming language such as Python, C++, and other popular coding languages. Nowadays, generative AI and LLMs such as ChatGPT, Claude, Grok, Llama, Gemini, and others are proficient enough that you can use the AI to generate code for you. In the same manner that you enter a prompt into AI to generate a story or answer your questions, the AI can instead produce programming code on your behalf. An entire Python program, for example, can be produced in a few seconds or minutes, instead of handwriting code that might have taken hours, days, or weeks to devise.

This might seem to be a dour sign for software engineers. A non-tech user can presumably write a prompt in natural language that asks the AI to write a specialized program to do this or that, and the user sits back while the AI does all the heavy lifting. Voila, a program is produced and ready to be run. Easy-peasy. No programming skills required. Just compose a suitable prompt that will spew out a shiny new, ready-to-go computer program. Nice.

There is a hitch right now.

The difficulty is that AI isn’t fully adept at producing programming code. There is a notable chance that the code will have some number of errors. Another notable chance is that bugs will be inadvertently incorporated into the code. Even worse is that a devious behind-the-scenes evildoer might have planted nefarious elements inside the AI so that when the code is generated, there are hacks included, and the code is going to perform evil acts once you begin to make use of it.

For the time being, software engineers can breathe a bit of relief that possessing honed software development skills is still highly prized. The relief might be relatively short-lived. Ongoing advances in AI could radically increase the code generation facility and ultimately do as well or even better than human software developers.

Software engineers need to be watching their backs.

Vibe Coding Hangover

Anyone who has been immersed in the computer field for a long enough time is fully aware of the many attempts to allow end-users to write computer programs. This is an endless quest. And it has generally been an overall ongoing disappointment. But the latest craze of vibe coding has some adamantly claiming that, despite prior efforts which have had mixed results, we have finally landed on the proper solution, namely letting AI produce your source code.

If the AI did a perfect job of doing so, we might be venturing somewhat into rather uplifting territory. The code, though, is currently often replete with disconcerting troubles. The troubles must be seriously dealt with. You cannot safely use generated code that then proceeds to work incorrectly or possibly goes hog-wild into performing evil acts. That’s obviously not good.

Who or what will take care of the improperly generated code?

Vibe Coders Not The Right Choice

I’m sure your first thought is that the vibe coder ought to do so. They generated the code; therefore, they should be responsible for perfecting the code. Period, end of story.

A big challenge is that the very same end-user who produced the code via writing prompts might not have any clue about programming and what the heck is going on in the code that has been produced. It is utterly a black box as far as they are concerned. An analogy would be someone who knows how to drive a car but is clueless about how to fix the engine when it conks out on them. They lift the hood, and everything looks mysterious and baffling.

I want to point out that if a vibe coder happens to also be a proficient software builder, they likely can indeed look at the generated code to fix it. Ergo, that’s a circumstance of the vibe coder doing both the code generation via prompts and then doing the debugging on the generated code.

The thing is that vibe coding is presumably supposed to be a widely adaptable approach to producing programs. A vital assumption is that end-users having near-zero programming knowledge will use AI to produce programs. The widest possible use of vibe coding will be by people who aren’t going to reasonably be expected to tackle the code that has been generated.

Getting AI To Clean The Code

A notable assertion is that if AI generates foul code, tell the AI to get its act together and fix the darned stuff. Just keep imploring the AI until the code becomes pristine. This might take a series of browbeating prompts, but that’s not a big deal. An end-user can simply command the AI to do a cleanup, wait and see what happens, and if needed, go ahead and repeat the declaration doggedly.

Sorry to say that this repetitive act of AI-based code cleanup has problems.

Sometimes the AI does this flawlessly, and the code is turned into pure gold. I wouldn’t bet on it. Any relatively complex code-generated program is not going to lend itself to easy fixes. I’m not saying the AI won’t always be unable to clean up the code. Sometimes it does, sometimes it doesn’t. Once again, the future suggests that the AI will be improved such that the code is perfect at the get-go, or the AI is proficient at squaring away the code that has been generated.

The disturbing aspect is that the AI can introduce new problems while in the quest to do the clean-up. Suppose the AI cleans up some of the code. Great, wonderful news. Meanwhile, the AI introduces several new bugs and errors. Whether you will even be aware that the AI has additionally messed up the code is probably not in the cards. The AI is bound to tell you the opposite, proclaiming that the world is a happy place and all your code is good to go. Don’t believe the AI.

The zillion-dollar question then arises as to how else the code can be fixed?

Humans Come To The Rescue

Here’s what we can do about this dilemma.

Whenever a vibe coder produces code, they are presented with an option to send the code to a specialist who is versed in vibe code cleanup. This is a particular specialty. The specialist has a deep understanding of how AI tends to generate code. They have strong skills as seasoned software developers and have likely been a legacy software maintenance crew member for a while.

The cleanup specialist inspects the code. A series of tests is run. Whatever seems to be wrong is ferreted out. Once the specialist has a semblance of where the issues are, they proceed to change some of the code. They might add additional code. Some code might be deleted. The aim is to turn the code into a properly functioning program. Plus, they shrewdly make sure to catch and remove any evildoing portions.

Getting The Job Done

Right now, this kind of cleanup is being done mainly on an ad hoc basis, such that you might happen to know a software developer who will kindly look at your AI-generated code and do their magic on the code. Some enterprising software developers are hanging out a shingle offering their services as a cleanup specialist.

The expectation is that entire farms of cleanup specialists will be formed. When an end-user is doing vibe coding, they will be presented with an option to route the generated code to a cleanup farm. Costs come into the picture. Do you want a quick-and-dirty cleanup, or a completely exhaustive cleanup? How much is the cleanup worth to you? Are you willing to wait to get the code cleaned up, or are you in a hurry?

And so on.

Software developers who are looking for a few bucks on the side are willing to take on these duties. Startups might be formed to do this as a service. Possibly existing programming shops will see this as a tempting new market and decide to jump in with both feet.

When there is a buck to be made, a rapidly emerging necessity can be met with ardent earnestness.

Problems Aplenty

There is no free lunch in the cleaning up of AI-generated code.

First, the cleanup by human hands doesn’t guarantee perfection. No developer in their right mind is going to guarantee that your AI-generated code will be spotless. If anyone does make that promise, I’d bet that it is some kind of sneaky setup or maybe a shady scam. Be careful. You could be taken advantage of. Suppose the Byzantine code comes back unclean. You don’t realize this has happened. Bad news for you.

Second, cleaning up the code now locks you into a one-off. Here’s the deal. Once the code has been changed by the cleanup effort, you have code that is stuck in time. If you go back to AI and ask it to re-generate the program, there is no ready means to somehow integrate the cleanup of the prior version. You will need to do a new clean-up.

Imagine then that you realize after the first cleanup that your prompt was amiss. The code doesn’t contain the needed functionality. You are between a rock and a hard place. Either keep the changed code as is or ask the software developer to now enhance the code to do new things. Your desire to get the AI to generate a new program that has a greater range of what you wanted will be out of step with that other code. You will need to get out your checkbook and pay for an entirely fresh cleanup on the latest generated AI code.

It’s a great business model for those providing the cleanup services.

Third, the cost of getting the code cleaned up might exceed having contracted to have the code crafted by hand in the first place. Sad face. The possibility exists that if you had paid a software developer to write the program, it would have been less expensive than trying to go the “easy route” of entering prompts and generating code. The eagerness to do vibe coding has to be balanced against the issues associated with the generated code and the risks, plus associated costs, of dealing with the aftermath.

Sloppy Vibe Coders

There are a lot more trials and tribulations on these sobering matters.

If vibe coders know that they can rely on cleanup specialists to hopefully turn the generated code into something runnable and proper, will vibe coders opt to let their guard down and be sloppy about prompting AI to generate the code?

The saga goes like this. A vibe coder might be cognizant that they should mention in their prompts a variety of important points about the AI generating good code. There is a definite need to give the AI overall guidance about the code generation. Without any indication of what to do, the chances of the AI generating sour code are generally increased.

But vibe coders might gradually land into a mental zone that the alleged failsafe of the human-based cleanup is going to get them out of any troubles. No need to rant and rave at the AI about its code generation. Let the AI do its thing. The wise-owl software engineer working downstream will do the cleanup.

A death spiral associated with vibe coding ensues as the AI-generated code worsens, and the effort by the cleanup crew becomes equally complicated and costly. A double whammy ensues.

The Future Is Somewhat Cloudy

What kind of career progression is possible if a software developer decides to dip into being a vibe coding cleanup specialist?

The work itself is not necessarily going to be especially intrinsically rewarding. Programmers who develop code from scratch usually get both glory and big bucks. Dealing with the maintenance of code or fixing code is often considered a less glamorous task. In the case of AI-generated code, there is an insider debate about which is worse: maintaining or cleaning up human-generated code or AI-generated code.

It’s a toss-up, for sure.

Does having the glorious title of cleanup specialist help your software development career or dampen it?

The answer is that it partially depends on what the future holds. Suppose that down the road, a lot of respect is garnered for those software developers who stepped up to the plate to do this kind of work. It might be a plus on their resume. A badge of honor is earned.

Ominous Clouds

On the other hand, skeptics and cynics tend to be insistent right now that the clean-up activity is not going to last very long. AI is going to soon enough do either better code generation or be better at fixing AI-generated code (or both improvements occur). As such, anyone who took this type of interim role is going to be out of a job and possibly have something on their resume that will be a puzzlement.

Why would it be puzzling?

Because later, just about everyone will have forgotten about that quirky moment in time where humans were asked to clean up AI-generated code. It will be nothing more than a quaint memory. You will get quizzical looks and be asked to clarify what you did in this hazy, here-today, gone-tomorrow role. No meritorious credit will be granted. The perception will be that you took an oddball side tangent on a normal career path.

Is there enough money to be made so that you can cash out on the vibe coding cleanup bandwagon and not worry about what happens further down the road?

Nobody can say. A flash in the pan might be happening. On the other hand, perhaps this is one of those rare and unforeseen opportunities that give rise to the proverb of no guts, no glory. The other classic adage that comes to mind is that nothing ventured, nothing gained.

A final thought for now.

As Mark Twain cleverly noted: “The future interests me — I’m going to spend the rest of my life there.” If you are a software developer, builder, engineer, or an all-around coding guru, you might give some in-depth consideration to becoming a vibe coding cleanup specialist.

Please only proceed in that direction if it strongly gives you a profound sense of positive vibes. Good, good, good vibrations ought to be at hand.

Exit mobile version