A Stalemate Between Software Jobs VS AI

Source: Dev.to

AI innovation – an unstoppable force – meets an immovable object – software jobs. So many arguments, endless predictions, countless hours wasted. Neither side is winning.

If you’ve read my previous articles and some of my tweets, I am very skeptical of AI and do not believe software jobs will be replaced by it.

However, the goal of this article is not to win an argument, prove a point, or filter through the noise. I want to listen to the noise, find where it’s coming from, and study why it’s being made.

Idle Observers

I will start by locating where the noise is coming from.

The majority of people who take part in the argument about AI and software jobs are what I refer to as idle observers – people who cheer for their side without taking any concrete action to back their beliefs.

• Those who hold that software will not be replaced by AI are just sitting ducks, waiting for the so‑called “AI bubble” to pop. Sitting idle and hoping to outlast the enemy without taking any action is not winning. The AI hype train has dragged on long enough to safely assume there will be no rapture for those who think they know how to code.
• On the other side, people who heavily use vibe‑coding tools for dopamine and clout while boldly making unsupported claims aren’t doing much either.

Binding Factors

Now that we know where the noise is coming from, let’s talk about why it is being made.

In the ocean of opinions on this topic, I’ve found that the two primary factors are Value and Risk.

Value

Many people don’t realize that “value” is subjective. In the same sense that a diamond offers little to no value to a starving animal, the best sorting‑algorithm implementation offers little to no value to the average end user.

Since the argument about AI potentially replacing jobs revolves around the rapid generation of software projects, it’s worth examining the value of software.

The Value of Software

• In software, incompatibilities across operating systems, CPU architectures, and web browsers limit value. A Windows‑only application is of no value to a Mac user, and vice‑versa.
• A content creator may find specific video‑editing software valuable enough to (God forbid) pay Adobe for it, while others look for alternatives or don’t edit videos at all.
• If a software engineer discovers a tool that drastically improves his workflow, he probably won’t tell his grandmother (or any non‑technical person) about it because he knows it will be of no value to her.

AI tools are extremely versatile and can provide immediate value to most individuals, especially when it comes to generating code. Because AI tools can fit many users’ needs closely, people who use them are more likely to share them with others. Sharing potentially helpful things is an act of kindness, not necessarily a bad thing—unless the thing is an AI‑generated slop video.

In conclusion, the hype around AI exists because people find value in it.

Risk

Risk pulls in the opposite direction of Value when it comes to AI‑generated code and is primarily what keeps software jobs rooted in place.

• Both humans and AI are unpredictable. An employee could randomly steal workplace equipment or commit fraud. Likewise, an LLM may hallucinate or refuse to provide harmful information.
• Most people agree that hiring an experienced worker results in less risk than hiring an intern. Experienced workers are paid more because of their ability and the lower chance of risk.

Risk cannot be measured precisely, but it can be reduced with incentives.

Incentives

A soldier defending his country is more likely to fight harder because he has something to fight for. Similarly, employees work because of the incentive to get paid. The reason most people don’t play life‑threatening practical jokes in the workplace or steal equipment is that they want to continue receiving pay.

Unlike employed humans, LLMs have nothing to lose. No matter how aggressively you shape their instructions, they have no real incentives to perform better. This mirrors the argument that communism fails: if someone’s basic needs are met regardless of contribution, there is little motivation to contribute.

Conclusion

I hope this article gives you a better understanding of both sides and a topological view of how silly we are when we butt heads while accomplishing nothing.

I want to stop fighting about AI and software jobs. Let’s be at peace with each other while we build great things.

(This does not apply to Communists. I will not be at peace with Communists.)