Gen Z programmers, breathe a sigh of relief. Millennials devs, it’s all good. Your Gen X & Boomer elders have seen this “X technology will replace programmers” before. Jesse Warden has got you.
1959: COBOL was called “Common Business Language” so non-programmers can read it. We were so expensive, the Pentagon got involved. Now COBOL Cowboys charge up to $1,000 per hour. The syntax morphed to OOP to be more likeable/usable by programmers.
197?: MYCIN, the 1st AI/LLM like inference engine, used to recommend antibiotics for infections. Despite high scores, wasn’t used because of hardware limitations (like the data center bulid out & RAM shortage drama we have today). Later Intellicorps LiveModel allowed you to leverage this AI to model business processes. This in turn spawned many data scientist & software consulting opportunities.
1973: Wang 2200 computer was called “computing calculator” so it wouldn’t scare customers. Marketed as “exceptionally easy to operate” and “no esoteric skills are required”. Yet saving files was so obtuse, many users would reach out to tech support (reminds me if iPhone file saving). You had to program it in BASIC vs. “File > Save As or Command + Shift + F”.
1982: OpenEdge Advanced Business Language: programming language w/integrated database build for people who weren’t computer scientists. Same later “I guess only programmers use this, let’s make it more OOP”.
1983: IBM’s Query Management Facility & Crystal Reports ensured business people could get reports without IT involvement. Some of my early consulting work was “customizing these charts for this data with these systems”.
1985: Went the other way, attempting to remove the Domain Expert via CLIPS (NASA’s AI Language). Some used these Expert Systems to generate code which sometimes was better than a dev could write. You don’t hear much about expert systems from the 90’s – 2000’s, but you DO hear about devs building rules engines leveraging those Domain Expert’s and their knowledge.
1991: Visual Basic, PowerBuilder, and Delphi lead the way in “using visuals, you don’t have to know much about code and databases”. Then those apps reached a level of complexity, consultants/freelancer devs were hired to fix, modify, or rewrite these apps.
1994: Many CASE tools like Rational Rose utilized visual boxes and arrows to represent domain models, such as UML, and these would generate your application, not needing developers. UML’s creators today use UML to communicate architecture to other devs, but doesn’t use it to generate the code.
2000’s: Business Rules Engines were touted to allow Business Analysts to make decisions without programmers (offering a loan, credit card, or mortgage). Many data specifications, tools, libraries, and microservices are created to allow access to these rules, and today many devs & data scientists build these systems to run businesses.
Excel: … yes, the business doesn’t need us here. All hail the almighty Excel.
With the exception of Excel, all of these innovations created to replace devs either led to more of us needed or new job roles created. All lead to new tools and approaches. COBOL to using English vs math in coding syntax, MYCIN for LISP & Prolog to be cool, inference engines to be a thing, including Bayesian networks, RAG, logic programming, and truth engines.
Wang + others lead to the personal computer market, word processors, & accessible accounting tools to run businesses & people’s lives. PC’s were gateway drugs for many people to be exposed to BASIC, VB, and game dev. The UI & case tools of the 90’s raised the bar on expectations of developers for IDE’s; we expect IDE’s nowadays to visually help us navigate our code intelligently and allow the LLMS to help.
The RAD Tooling was overshadowed by the RAD development style which helped influence the world away from Waterfall to more Agile styles of working. This allows us to continually learn with our users to help inform how to build things, leading to constant learning and fun. UML and it’s ilk has lead to all kinds of great data formats like JSON, Markdown, and Mermaid charts.
All of these are cool, new things we can learn to make our jobs better, or new types of jobs with software. They always try to replace us. We always end up with new toys instead.
So don’t worry about your software job being replaced, or no new junior positions. If you like being here because you have a short attention span, love to learn new things, and build stuff… or all 3, you’ve chosen a technically great time to get into software dev.
We’re all still figuring this AI thing out, which is fun! Part of software dev is bravely/excitedly wading head first into ambiguity and going “I wonder how this works… what does that button do…” While opportunities right now suck because of the economy & over hiring, there ARE a lot of people who really don’t want to be here who are leaving, and the economy historically always gets better. Just keep going.
Get that STEM degree, go to that bootcamp, or keep trying to break into your junior role; the amount of software we have to build is only increasing, and we’ll use AI/LLM’s to help us do it. Find your passion in the field, whether that’s web dev like me, game dev, systems programming, data science/machine learning, cyber… or maybe you just gravitate towards a particular language, that’s cool too. There are so many fun niches in software dev.
Leave a Reply