Back to the future
Published:
Recently, I watched the movie “Hidden Figures.” One thing that caught my attention was the fact that the advent of the IBM mainframe computer was a major disruption to the careers of human computers. The best way for most of them to survive was to transition to the role of computer programmers by learning Fortran.
The advent of coding agents like Claude Code, Qwen Code, Cursor, Windsurf, and Antigravity seems to pose a similar threat to the careers of computer programmers. Hence, to me, it feels like history is repeating itself. The era of the coder is evolving just as the human computer did. It is now time for programmers to pivot toward roles that go beyond writing code and instead focus on the high-level problem-solving and logic seen in competitive programming. So, I asked Google Gemini to list similar professions so that we can learn from history, especially the kind of pivots that worked.
| Tool | Extinct Role | Evolved Role | Pivot |
|---|---|---|---|
| Mainframe computer | Human Computer | Software Programmer | From performing calculations to writing the logic that calculates. |
| Digital Desktop Publishing | Linotype Operator | UX/UI Designer / Typographer | From physically casting lead type to designing digital layouts for readability. |
| Automated Switching / VOIP | Telegrapher | Network Engineer | From translating Morse code to architecting the packet-switching systems of the internet. |
| GPS / Satellite Tech* | Navigator (Sextant) | Geospatial Analyst | From calculating position manually to managing real-time satellite data streams. |
*Quantum navigation is replacing GPS.
On the other hand, the rise of Google Search revolutionized information access, fundamentally transforming the role of the traditional librarian. While physical foot traffic for basic fact-finding declined as knowledge became digitized, the profession evolved into Library and Information Science (LIS), focusing on digital curation and information literacy. In the corporate world, this analytical shift led many into Data Science. However, as generative AI tools like Gemini, ChatGPT, and Perplexity begin to automate routine data cleaning and statistical testing, the data scientist’s role is shifting once again. To remain indispensable, they must pivot beyond basic execution toward advanced statistical validation, algorithmic oversight, and the complex interpretation of AI-generated insights.
In 2024, the Nobel Prizes in Physics and Chemistry were both awarded for breakthroughs directly tied to Artificial Intelligence. John Hopfield and Geoffrey Hinton received the Physics prize for the foundational neural network discoveries that power modern AI, while the Chemistry prize was awarded to the creators of AlphaFold for using AI to solve the 50-year-old challenge of protein structure prediction. This shift signals that AI has transitioned from a niche experiment to an essential scientific instrument. Consequently, researchers who fail to integrate these tools into their workflows risk becoming obsolete in an increasingly automated landscape.
As a mathematician, I also feel the pressure mounting. For example, celebrated mathematicians like Terence Tao, Jordan Ellenberg, and Ken Ono have jumped headfirst into AI-assisted theorem proving using Lean formalization. However, history suggest this maybe an evolution rather tan an end. Chess engines became training grounds for chess players instead of killing the sport. Similarly, it will be interesting to see the impact of AI agents on mathematics competitions and research.