The original post was written in late 2023, edited and changed in the first months of 2024, but never published, because I was persuaded by some peers that had glimpse into it - too dark and negative. 2 years later and here it is - I published it this as is (only with some small formatting changes). Now, with a bit of pride that I got most thing right (in these 2 years). I plan on writing part two, with more predictions on how the changes might affect society and what are the possible end results (in more details).

Intro

I have a bachelor’s and master’s degree in computer science, specializing in intelligent information systems. I have 12 years of experience as a software engineer – started with Web, but quickly switched to the then shiny new(ish) thing: Android Development. I’ve been on different points in the software engineering process, down in the trenches and outside overlooking the deadline release battles. I’ve worked at startups, digital agencies, product based large companies – from “it doesn’t matter – let’s ship it” to “the name of this variable in your unit test is not intuitive enough”. I’ve worked on passion projects, blogging, published a somewhat popular Android library, I’ve doing data science in Python and trying to center a div in CSS. I consider myself a valuable addition to a software engineering team – knowledge, experience, creativity and some bad jokes on stand ups. I’m humble enough to know that I’m replicable – as a software engineer or researcher, a very strong and genuine “will be missed”, but still, replicable. Looking in the eyes of the upcoming AI development automation, we, software engineers, will all be humbled and sooner or later will come to the same realization.

Post pandemic tech market was booming, the abundance of job opportunities was overflowing the tech fountain of career perspectives. This has resulted in inflating tech employment and creation of a big cousin of software engineer’s comfort, who has been able to indulge in conspicuous consumption more than ever. Everything started to shift with global political events which was aided by the introduction of ChatGPT. Ever since that moment there has been progress in tooling that erodes, bit by bit, the rock of software engineers’ responsibilities, narrowing down the job description. The aim of the AI iterations is to reach a point where it can do everything a human would do, without emotions and imperfections. The end goal is to create a know-it-all entity also referred to as artificial general intelligence (AGI).

If you look into the software engineering corner of the Internets, the theme playing out loud is “will I be replaced by AI?” – and I’m going to try to answer that and make a prediction. Hopefully a good one. Drumroll, the answer is: yes. Let me elaborate.

Predictions

My predictions are going to be centered around tech and how those changes will affect society. To be noted, there’s no branch of the economy that will not be affected in a substantial manner.

1) The changes will happen in waves of invention and adoption – even if the invention moves exponentially the adoption will slow down the pace. How much remains to be seen.

First wave. We’re currently in the first wave, where LLMs are packaged in the form of tools that are good enough to aid the developer in the daily work. Some very repetitive parts of the engineering process are easily replicable with the tools being available. Mainly, it’s just an addon to what developers already do, but helps with productivity and brings value to the shareholders. Jobs like translators, content writers, manual QAs can be rendered obsolete, but are not. The industry sees a shift and is reluctant for new investments into hiring people, rather than developing these tools, or waiting for these tools to be developed. Minor adaptations already have begun, but very few companies have official and firm stances. Data privacy is still not resolved, one of the many reasons why companies are hesitant to start applying the new tech on mass. This is where the people that say “you’re most likely not going to lose your job (because of AI)” are right. We have 1–2 years left from the first wave.

Second wave. Tools like Devin and more to come (AutoDev from Microsoft?) become good enough to replace junior devs and do most of the work. Data privacy is resolved, and more companies start to embrace the “AI dev” tools. First AI Dev digital agencies appear – offering packages to companies for “subscribing” to their dedicated AI team of devs. Very few new openings are going to be posted, low demand, high supply. Layoffs are going to have their moment – each next round will be higher than the previous one. This won’t be happening just in tech, but also in most data-driven industries. The new tech is going to produce tools to replace other job descriptions, while it also eats itself – a truly poetic end. This is when you’d be really worried about your job if you still had one. Getting a job once you got laid off will be harder than ever, close to impossible. The economy shifts are going to be apparent and first “organic” software movements are going to appear. The pushback from developers (and other workers that were laid off) is going to take some form. A lot of talent is going to be unemployed and that talent and experience, with the help of the very same AI that got them in that position, will be able to compete against the now very big companies. If there’s no legal intervention and most of the AI breakthroughs stay open, big tech is going to be under threat in terms of monopoly. Building complex systems will be easier than ever, an opportunity that a sizable chunk of the unemployed devs will take. I expect this wave to last 2025–2035.

Third wave. This is going to be a period that is mostly adaptation and social amortization of the automation consequences. How societies are going to handle the second wave will determine what will happen in the third one and what pace AI adoption will have. Tech is going to be focused on displacing all other jobs that are now considered safe – manual labor jobs that need more context and require multiple models to cooperate. With advancements in robotics this will be achieved. Home robotic appliances for ironing shirts and folding clothes will be good enough that no human assistance will be needed. It’s unlikely, and I’d be very surprised, if by this time we achieve AGI, but some pretty good forms of it are going to be out there. I expect this period to be 2035 until a viable AGI is introduced.

2) NO, AI won’t “create more jobs than it’ll displace”.

The “new jobs” that are going to be created will be narrow, harder for novice developers to do and/or to be seen as eligible by companies. The price of labor will drop, and tech won’t be an attractive field to get into.

Social consequences of automation

Society is built on norms and rules that make the system run. The massive introduction of AI will challenge most of the norms and all the rules that have built modern societies. The very basic rules of setting up the social hierarchy, climbing the social ladder, through actions of acquiring resources because of work, will be diminished or totally nullified. The stability of the system depends on how fast this change is going to happen and how it will be handled.

Almost certainly we’re going to see economic turbulence in the short run, followed by very probable social unrest and some dystopian form of new society as the end result.

Most likely outcome will be frozen or very inert social mobility, two main classes:

• The "organic" oligarch/winners of AI revolution class. They’d be able to enjoy goods and services developed and produced by people, not because they are better, on the contrary because they can afford their costliness. Form of status signaling. Art, movies, food, barbers, designers - everything human made.
• The rest, the synthetic class. The consumers of the "free" products of AI - basic goods and services enough to live lives happy enough not to rebel or revolt and not good enough to afford the "organic" luxury.