How artificial intelligence is changing the economic structure?

With the emergence of LLM models, some people have new opportunities they didn't even dream of. For others, it's the risk of losing their job or getting a lower salary.
3 min

With the emergence of LLM models, some people have new opportunities they didn’t even dream of. For others, it’s the risk of losing their job or getting a lower salary.

As always when new technologies appear:

hen trains first appeared many believed the human body could not withstand traveling faster than 30 mph. Some doctors warned that high speeds could drive people insane. 😌
When electricity emerged, it was called the “devil’s invention”. Many thought electric light could blind people, and so on.

In general, people don’t like to leave their “stable” worldview. This is both bad and good for those open to new opportunities.

Well, what does the future really hold for humanity? 🤔

According to research, 80% of all professions in the US will change because of AI. In each of them, at least 10% of tasks will be automated with AI without quality loss. For 19% of all employees, automation will affect over 50% of daily tasks.


The higher the education and salary level, the more automation will appear in the profession. For a barista, plumber, or hairdresser it’s almost zero. For a Wall Street analyst or legal counsel – very high. For professions with an average salary of $30k/year, the share of tasks subject to automation is about 5%. For salaries above $80k – it’s 50%.

Overall, given the current state of technology, the authors expect automation of ~15% of all tasks in the economy without quality loss. And considering new AI-based products, this figure could reach 47%.

According to the authors, LLM models are a general-purpose technology like electricity. Most use cases have not yet emerged or are in the very early stages. Over time, the technology will improve and these numbers will change.

I’m personally in several groups where programmers, analysts, and just interested people create their models and share discoveries, tools, questions, and answers.
It all looks like solving a Rubik’s Cube. In these groups, there is constant activity and creation (both interesting and not). And we haven’t heard about the pros and teams working on their own projects yet. Some are now emerging from the shadows, but many projects are still behind the scenes. I can’t even imagine what will happen not in 5 years, but next year.

I personally see tremendous opportunities in this technology to expand my capabilities and tools.

LLM is a trend, and I see many people afraid of “not keeping up”. I want to tell such people: to calm down and focus not on catching the trend but on your work and tasks. Trends are based on people’s reactions, and people slowly adopt new tools. Perception of new trends is uneven – first enthusiasm, then disappointment, then habituation. These phases alternate like waves.

Everyone will be able to get what they want at their own pace.