TL;DR
A long-time programmer argues that every past wave of tooling that promised to eliminate developers instead expanded software and demand for programmers. He says large language models (LLMs) are different in scale but not in kind: they often introduce reliability and productivity problems and cannot replace the human skill of turning vague human needs into precise computational designs.
What happened
The piece is written by a programmer with 43 years' experience who reviews repeated cycles of technology that promised to make software developers obsolete — from early compilers and third-/fourth-/fifth-generation languages to Visual Basic, wizards, executable UML, and modern no-code/low-code platforms. The author says those predictions have consistently been wrong: each wave produced more software and more developers, a pattern he frames as an example of Jevons Paradox. Large language models are now receiving similar doomsday claims, but the author contends they differ mainly in hype and scale. In practice, he argues, LLMs often slow many teams and create less reliable, harder-to-maintain code unless underlying process bottlenecks have already been solved. He also links recent workforce changes mainly to economic factors — pandemic over-hiring, higher borrowing costs, and heavy infrastructure spending on data centers — rather than to AI eliminating developer roles. The author expects modest AI assistants to help with prototypes and inline completions, but insists that human developers will remain in charge when it matters.
Why it matters
- Historical cycles suggest new tooling increases software output and continuing demand for developers, not their elimination.
- LLMs can introduce productivity and reliability challenges for many teams unless process and skills gaps are addressed.
- Economic forces, not AI alone, have been a primary driver of recent hiring and layoff trends in the industry.
- Organizations should consider hiring and training developers now to avoid talent shortages when the hype subsides.
Key facts
- Author reports 43 years of experience as a computer programmer.
- Past technologies touted to replace programmers include Visual Basic, wizards/macros, executable UML, and no-code/low-code platforms.
- Earlier generations (3GL, 4GL, 5GL) and early compilers were also once claimed to end the need for programmers.
- The author characterizes the historical outcome as more programs and more programmers, citing Jevons Paradox and a $1.5 trillion-per-year example.
- Large language models are louder in public debate but, according to the author, frequently slow many teams and reduce software reliability.
- The hardest part of programming remains converting ambiguous human thinking into precise, unambiguous computational designs.
- The article says there is no credible evidence that AI is replacing software developers at scale; recent layoffs are tied to over-hiring, rising borrowing costs, and heavy data-center investment.
- Hyper-scale LLMs are costly to build and running them may not be a viable long-term model, the author suggests.
- Likely near-term uses for AI in development include prototype generation and inline completion rather than full system replacement.
What to watch next
- Whether teams using LLMs see net productivity gains or further slowdowns as adoption scales.
- How hiring and training activity changes once organizations reassess AI-driven productivity claims.
- Viability and business models of hyper-scale LLMs over the coming years.
- not confirmed in the source: precise timeline for arrival of artificial general intelligence (AGI) or its impact on developer roles.
Quick glossary
- Large Language Model (LLM): A machine learning model trained on large text datasets to generate or predict human-like language, often used for code generation and text tasks.
- No-Code / Low-Code: Platforms that let users build applications through visual interfaces or simplified tools with minimal hand-written programming.
- Jevons Paradox: An economic concept where increases in efficiency or productivity can lead to greater overall consumption of a resource or service.
- AGI (Artificial General Intelligence): A hypothetical form of AI that can understand, learn, and apply intelligence across a broad range of tasks at human-level capability or beyond.
- Compiler: A program that translates source code written in a programming language into executable machine code or another target language.
Reader FAQ
Will AI replace software developers?
The source says there is no credible evidence AI is replacing developers at scale; economic factors explain recent workforce shifts more than AI-driven substitution.
Are LLMs making software development faster and more reliable?
According to the article, for many teams LLMs currently slow development and produce less reliable, harder-to-maintain code unless existing process bottlenecks are solved.
Should companies pause hiring developers because of AI?
The author argues companies should not pause hiring and suggests hiring now to avoid shortages once the hype abates.
Is artificial general intelligence imminent?
not confirmed in the source
The Future of Software Development is Software Developers I’ve been a computer programmer all-told for 43 years. That’s more than half the entire history of electronic programmable computers. In that time,…
Sources
- The future of software development is software developers
- How Software Development is Changing Forever, and How …
- The Future of Software Development: Why AI Won't …
- Software Engineering Isn't Dead: The Bright Future of …
Related posts
- Why the Internet Is Bad for Democracy – A 2005 ACM Paper Reviewed
- Ford Reiterates Ongoing Support for Apple CarPlay in Future Vehicles
- Left Behind: Futurist Fetishists, Prepping and Earth’s Abandonment