Office work or manual labor: Who will 'survive' the AI revolution?
In the near future, artificial intelligence (AI) can replace humans in many jobs, from writing reports, analyzing data to driving and carrying goods. This raises a big question, will office work or manual labor be 'safer' from the intrusion of AI?
Millions of workers have continued to work in factories for years, while AI has rapidly replaced novice programmers thanks to breakthroughs in large language model (LLM) technology.
Meta CEO Mark Zuckerberg recently announced that the company will begin cutting mid-level software engineers as early as this year.
Similarly, the CEO of Replit – a platform that provides website building tools, asserts that they no longer need professional programmers thanks to their AI product, which is capable of writing software applications just from user prompts.

A number of other companies are also developing similar solutions, helping users create websites and applications without programming knowledge.
Meanwhile, manual labor jobs like manufacturing and construction are being automated at a significantly slower rate.
So, will the LLM boom prolong the life of blue-collar jobs, or will robots soon catch up? Are programmers and office workers really in danger of disappearing?
AI will not replace software engineers
Devansh Agarwal, a senior machine learning engineer at Amazon and author of several papers on AI, believes that AI will not replace software engineers. Instead, AI will play a strong supporting role, helping them reduce repetitive and boring tasks.
Software engineers will remain in demand due to the complex nature of the job, which includes tasks such as maintaining and reviewing code, designing technical documentation, building algorithms, and troubleshooting in production environments, he said.
Agarwal points out that even the largest language models, with more than 600 billion parameters, are still not able to accurately perform a specific task. He is skeptical that even if models increase the number of parameters by 10 times, they will not be able to completely replace software engineers, due to the limited data available for training.
“The problem is much more complicated. To get good results from LLM, we have to experiment multiple times with different prompts, which significantly increases the cost. Furthermore, it is difficult to convey the complete product requirements to LLM and expect it to generate a complete solution, as these requirements are always changing,” Agarwal explains.
He predicts that the LLM will have the same impact on programming that software engineers have had in the past. However, current market trends show that the demand for jobs that simply involve writing code is decreasing, signaling that programmers will need to expand their skills to adapt to the future.
AI has a big impact on theprogrammerlow level
AI will have the biggest impact on entry-level positions in technical fields like coding and data analysis, said Mantas Lukauskas, head of AI technology at web hosting company Hostinger and a researcher at Kaunas University of Technology in Lithuania.
“Before, you needed to understand what to calculate, but now you just upload the data table to ChatGPT and it gives you all the information, you no longer need a junior data analyst,” Lukauskas notes.
A simple programming job like creating a website can be done by companies like Vercel, which allows one to code without any specialized skills.
The growth of programming tools and AI assistants like Github Copilot will require future programmers to learn new skills.
“There is creativity, then there is general computer knowledge, then there are AI-related skills. Critical thinking is also essential, many people rely too much on LLM. They create code and just copy it without checking if it is optimal,” said Lukauskas.
Beyond coding, AI is impacting many other roles. According to Lukauskas, Hostinger’s chatbots handle about 50% of the workload, while AI-based tools improve productivity by eliminating repetitive tasks and allowing humans to focus on more strategic work.
AI won't take programming jobs awaybecome obsolete
The Anthropic Economic Index, a report by the American technology company Anthropic, offers a glimpse into how AI is changing the labor market. The report is based on millions of anonymized conversations on Claude.ai, an AI assistant built by Anthropic.
It’s not just computing and mathematics where AI has the highest adoption rate (37%) – arts, design, entertainment, sports and media are also seeing a dramatic shift, with AI adoption rates reaching 10.3%. The study’s authors say this trend reflects the explosion of marketing, content creation and other forms of media.
Natalie Lambert, founder and managing partner at consulting firm GenEdge Consulting, points out that AI is now taking on many tasks in marketing, from content creation and basic design to data analysis and competitor research.

“In the near future, marketing teams could achieve similar performance with at least 50% fewer people,” she predicts.
As for the impact of AI on the programming field, Lambert cited her own experience as an example. She said that a business can now build a website using AI tools for as little as a monthly subscription fee, whereas previously they would have had to spend thousands of dollars to do the same.
Still, she insists AI won’t make programming obsolete. On the contrary, new tools will help developers focus on more complex, high-value tasks, like security or custom functionality.
“The challenge is that as the demand for senior developers increases, entry-level and intermediate developers may struggle to find opportunities. The important question is how to become a senior developer without starting from scratch?” she shared.
Programming skills will become important and necessary in many different fields.
Severine Zaslavski, Vice President of Global Product Development at the multinational corporation ManpowerGroup (USA), affirmed that programming is still a promising career choice for those new to the technology field.
“Those looking to get started should explore emerging areas like cloud development, edge computing, quantum computing, and specialized skills like ethical AI practices,” she recommends.
In addition, core skills such as analytical thinking, mathematical proficiency, adaptability and a desire to learn will be key to success in a rapidly changing job market.
ManpowerGroup predicts that over the next two decades, programming will increasingly become an interdisciplinary field, combining expertise from fields such as biology, neuroscience and many others to solve complex problems.
For knowledge-based jobs, Zaslavski says automation could reduce the need for back-office support and customer service. But rather than replacing them outright, AI will likely boost STEM, creative, and entrepreneurial industries by boosting productivity.
Anthropic's report also found that AI is largely being used to support and enhance human capabilities (57%), rather than completely automate work (43%).
Can robots catch up?human ability?
While AI is making significant changes in many office jobs, its impact on manual labor remains quite limited.
According to Anthropic's report, the occupations that use AI the least often involve manual labor, including transportation and material movement, health care support, agriculture, fishing, and forestry.
Lukauskas points out that, for years now, businesses have been ramping up their robot deployments, leading many to worry that a wave of automation will replace their jobs.
“However, since the emergence of large language models, the situation has changed. Although AI can influence decision-making in robotics, it cannot directly intervene in the physical world,” said Lukauskas.
Agarwal, an advisor at a robotics startup, has a different perspective. He believes that manual labor jobs are at a higher risk of being replaced than white-collar jobs, as the ability to program robots with generative AI (GenAI) becomes increasingly easy.
However, Zaslavski predicts that in the near future, jobs that require hands-on skills, physical presence, and human interaction, such as construction workers, delivery drivers, personal care assistants, and healthcare workers, will be less likely to be automated.
“These roles rely on adaptability and agility in a flexible work environment, which AI or robots are unlikely to completely replace. In addition, high-tech jobs such as renewable energy engineers or autonomous vehicle experts also have long-term prospects, as they fit into the trend of green energy development and future infrastructure,” said Ms. Zaslavski.