Since OpenAI's chatbot, ChatGPT, was released in 2022, people have been concerned that AI will take over jobs, especially as the technology advances rapidly. SWEs are using Claude Code; in 2026, you probably would not see anyone code without it. Linus Torvalds, the founder of Linux, said, "I don't understand this whole idiotic concept of 'don't use AI for X'." Radiologists are using AI to analyze medical images. Your mom, probably, asked AI what to cook tonight. So, what can be replaced by AI? The answer is clear, as you can notice the pattern: whatever is repetitive.
Machines removing redundant work is an old concept. One of the earliest successful applications of a machine learning approach was used to spot cardiovascular diseases. Machines were trained on labeled ECG patterns and made predictions. It optimizes the detection of diseases by removing the manual, redundant work of comparing results repeatedly. (Lindholm et al., 2022) This is similar to modern AI approaches in performing different tasks: massive datasets, then outputs. Tech companies famously quote that AI accelerates scientific and medical improvements, coding tasks, and workflow automation. For example, in October 2026, Anthropic announced early progress in accelerating drug development using AI (Anthropic, 2026). The way AI accelerates certain tasks, though, is by eliminating the need for humans to evaluate the data manually. The thinking processes, for example: "What do we want to achieve?" "What are the steps?" and abstract ideas remain difficult for AI to think about, as it is a generalization of data. Data and the interpretations in modern scientific research are notoriously redundant. Researchers often spend an exorbitant amount of time finding the correlations between x and y. Machines, however, can spot them significantly faster than humans. This applies to computer science. Coding is repetitive. The same formula can be used thousands of times. That is why developers used to go to Stack Overflow for solutions. The complex design, though, such as the architecture, requires human thinking. Therefore, AI is shifting the way we work: from looking at a hundred photos to knowing exactly how to act based on the instant results. That is the acceleration.
The way modern marketers frame AI is, to a degree, delusional. It cannot make you millions of dollars in a day. As mentioned earlier, the way it works is that it collects massive datasets (sometimes with labeled data; supervised learning), and then runs them through layers of algorithms, and then predicts what is going to happen. On a side note, LLMs are essentially predicting what the next token (word) would be; it is not inventing something new. That is why, despite AI being good at the heavy lifting, it is generally bad at making art and performing design-related work, specifically when it comes to expression and brainstorming. Typically, the artwork, video edits, whatnot, are generated from code, which is something AI is good at. But the results might be confusing because AI is not as good at presenting as a human does. It makes AI generic and hallucinates whenever it tries to explain something that does not exist in the datasets. When we understand how it works, then we look back at the delusional claims; we can identify what AI applications are workable and what is not. So, AI does make certain work more efficient as a tool; that is why every developer, or to an extent, everyone, uses AI, but there are fields where this tool is not applicable. They could be fields that involve creativity, decision-making, etc. AI-generated posters are one of the cases that go against creativity. It is understandable, as it removes the cost of hiring a graphic designer; however, it is not justifiable. AI will make the posters utterly similar, as, again, it is just data generalization. So, human-made graphics will eventually stand out. Not to mention the stupidity of AI stock trading. Even in quantitative finance, algorithms are essentially betting on tiny margins. “AI trading tools” that claim to profit in 2026, generally, are just a bunch of websites that scrape news and bet for you; you will lose money.
Now we know what AI is good at and bad at. Here, I am going to make my predictions. Given that AI is replacing repetitive work, and school materials in 2026 are still memorization-heavy, I believe the educational system will experience a great structural change in a span of ten years. Or it will be completely gone because participating in jobs that involve repetitive work after you get a degree is getting harder in the foreseeable future because of AI. We are, in fact, living in the best, or at least a whole different era of humanity compared to decades or centuries ago because of the technologies mankind has created. However, while AI is becoming increasingly significant in our daily lives, we have to understand its fundamental capabilities before dumping the responsibility for thinking and giving everything to AI, because thinking forward is the most valuable asset a person has.
Sources
Anthropic. (2026, October 27). Early progress on accelerating drug development with Claude. https://www.anthropic.com/research/early-progress-drug-development
Lindholm, A., Wahlström, N., Lindsten, F., & Schön, T. B. (2022). Machine learning: A first course for engineers and scientists. Cambridge University Press