Why is everyone trying to replace Software Engineers?

Economic Motives & Cost Cutting

• Many comments frame this as basic capitalism: software engineers are a large, expensive, “automatable” labor component, so companies naturally try to reduce or replace them.
• Even with high margins in software, some owners always want more; there’s no stable “enough profit” level.
• The AI-replacement narrative is also seen as a bargaining chip to suppress wages and tilt power further toward employers and shareholders.

Current Capabilities of AI vs Engineers

• Strong skepticism that today’s LLMs can replace real engineers, especially for:
  • Debugging complex production outages end-to-end.
  • Navigating large, messy, long-lived codebases and infrastructure.
• LLMs are described as useful for boilerplate, CRUD apps, tests, refactors, and greenfield work, but weak on nuanced systems and environment interaction.
• Some argue they’re still more “productivity tools” than “replacement tools,” especially for seniors; big value for juniors learning.

Future of the Profession & Junior Roles

• Several expect “code monkeys” and low-skill juniors to be displaced first; experts remain essential longer.
• Some foresee software maturing into a smaller, more elite, specialist field (like medicine or law), with fewer well-paid juniors and more unpaid/low-paid apprenticeship-style paths.
• Others predict that higher productivity will increase overall software demand and employment, but cut off the bottom rungs.

Industrialization of Coding

• Popular analogy: LLMs are like hydraulic cranes or die presses for code—fewer “blacksmiths,” more operators/maintainers and quality checkers.
• Counterpoint: code isn’t the finished product; the hard part is designing correct systems tied to real-world outcomes, where bad automation can be net-negative.

Comparisons: Offshoring, Low-Code, and Past Hype Cycles

• AI push is likened to offshoring: often promised savings, frequently disappointing due to communication/context issues.
• Similarities drawn to low-code/no-code and prior “developer replacement” waves that ended up increasing demand for good developers.
• Some argue AI is different this time and could reach or surpass typical developer capability within 10–20 years; others are unconvinced.

Power, Perception, and Who’s Really at Risk

• Several note that non-technical colleagues often don’t understand what engineers actually do beyond “typing code,” making replacement narratives easier to sell.
• Some argue AI is a bigger long-term threat to bloated management/coordination layers and large incumbent tech companies than to small, highly leveraged engineering teams.