I have been thinking for a while to write a post about the topic of AI and GPT and how that will influence or change the world we are living in. The reason I haven’t gotten into writing it before is that, like many other people, I have conflicting thoughts about his subject.
This will be a gloomy post or rant. Not sure what it is to be honest.
Most of the people my age, born last century, grew up reading science fiction that depicted a future where humans are free from the mundane - the premise being that robots and AI would take over these tasks and let us focus on intellectual endeavors. And then we grew up. The realities of the world are quite different which means that all of those fictional promises are just that - promises.
Why am I bringing this up? Well, we should be looking at the advent of AI-like systems, not from the perspective of “liberation from the mundane” but more of “<sarcasm>liberation from income</sarcasm>“. The world we built so far is focused on growth. That is the major metric entities within an economy are judged by. Growth is an impartial, non-empathetic concept with no regard for subjective human experience. Any possibility to improve this metric is seen as positive with no regard to any other impacted metrics such as the well-being of humans, animals or the environment. AI-like systems promise a lot of growth to the entities that want to make use of them. This growth will come primarily from cost reduction, by removing humans and accelerating product development**. These two avenues are intrinsically connected since, we, humans are the greatest bottleneck.
And yes, this already happened a few times before. And economists will swear that any time this displacement happened, we found other jobs to do. The biggest change was the proliferation of white-collar jobs. Most value-add businesses in developed economies are making use of those jobs. But what is next? If you displace people from those jobs, where can they go? This is a topic of great debate these days and I haven’t heard any convincing argument from any side yet.
So what’s next? Let me explore the topic a bit more.
Before I start on this topic, I feel obligated to point out that ChatGPT is not intelligent. It is a well-oiled machine that can generate text based on millions of text articles it has been trained on before. For this reason, various people tend not to take it seriously and even resort to calling this system a party trick. I agree and disagree with this interpretation. From my point of view, ChatGPT has shown us, that a lot of what human beings think is “intelligence” is not much different from a statistical correlation of information in our mind. This is a gross over-simplification but also it’s not. This phrase was not meant to be “smart” but rather to point out that a big chunk of what people consider intelligence, in reality, is not that special. Real intelligence is, of course, a lot further away from automation than people think, but, that doesn’t mean that it is required for most jobs.
The Changing Landscape
So, how many jobs require “real intelligence.” I would argue not that many. In an average business, you will find many people that do repetitive tasks that do not require any form of independent thinking beyond a simple “scripted model”. These are the first jobs to go. This has been happening for a while with less advanced technologies but will start accelerating in the next 5 years.
So which jobs are safe in the long term: There are no safe jobs in the “white-collar” world. Oh, the irony… we, information workers are replacing ourselves with a smile on our faces and a thank you note in our hands.
The hard jobs to automate are rare at this point. I am not an economist and will not pretend I know much in this regard but it will be interesting to see what the future brings.
I’ll come out and say it – I have never seen more appetite to replace one’s jobs as I see in the software building field. It is astounding how much focus we are putting into these projects. The call is “poetic justice” if you will. These changes will take time. Before that, we will all be very excited about the tools we are given; tools that will save us hours on writing code, debugging, understanding the context, etc. But slowly, the number of available positions on the market will start drying up.
If you are a young professional in the software field - I feel you. You will have a very uncertain future. One sad, and yet funny remark I’ve heard from several people is “I hope I will retire by the time this happens.”.
Should you go into a “machine learning” or “AI” field? Unlikely to make a difference. The AI builders are no different from other software engineers - they are also replaceable. Unless you are a researcher in a research institution/group, you will not be “safe”.
It’s About Numbers
If you follow the conversations on AI on the internet, you will find that people’s opinions differ wildly - from “it’s never going to happen” to “we’re doomed”. The former is obviously a self-defense mechanism from people too afraid to accept the inevitable changes and the latter is too dramatic.
The changes will come gradually as they have been for decades in other professions. The most likely scenario will see the number of jobs declining slowly in most white-collar professions. If it took 10 journalists to maintain a section of the paper last decade, it will take only 8 in the next 3 years, then only 5 and then, likely, just 1 or two. The same will happen with every other profession. How many mechanical engineers are needed to improve an engine? Not sure, but with an improving AI this number will go down - that is a certainty.
On a Positive Note
Let me start with a simple phrase: “It’s not that bad”. Jobs don’t exist in isolation, they are part of a greater cycle, which is part of the greater economy. What is the purpose of a business if not to make money? And how can you make money if you do not have customers? And customers only exist if they have income. It makes no economic sense to replace all the workers with automation since those workers are your potential customers.
This will require some changes at the societal level but there are reasons to be optimistic. On the other side, if the whole thing will collapse then it doesn’t make sense to worry about it now anyway.
And this is where I will stop! Cheerios!