AI and the bottom rung of the career ladder

LLMs can now write code 100x faster (or more). But only if it’s code in a greenfield project, or boilerplate, or a mixture of both. Creating the software is still hard, we’re seeing how vibe-coded software actually fails, with big companies that switched too early and too fast having famous breakages. Learning how to write the good code is hard, takes time, requires practice and guidance.

But, at the same time, because LLMs can output code 100x faster and executives are all captured by the siren song of GenAI, we’re seeing that jobs at the entry-level of the career are starting to disappear – in the short term, at least, but now more and more companies are focused on the bottom line and the short term only. Software craftmanship remains the differentiator between strong teams and companies who can build quality stuff, and everyone else. But, we’re cutting the bottom rung of the career ladder, the place where new entrants can learn the skills of the trade.

This is not new. History tells us of the time when crafts were practiced in guilds, when people would first enroll as an apprentice to a master, perform well, and only later be selected to inherit the business.

This all changed when the industrial revolution arrived. Craftsmen gave way to factories and the assembly line. Societies lost the slow, personalized skill transmission. We gained mass produced low cost Ikea furniture and personalized, sturdy stuff is hard to come by, the masters are very rare.

Over time, we adapted. There are still some places where the heir apparent is first taught the skills of the craft, step by step, over years. But most are inheriting the family business and then send it to the ditch.

We have a chance now to prevent a repeat. AI can still create new jobs and it’s in our interest to invest in newcomers, helping them learn the basics of the traditional software engineering practice at the same time as we all learn to use AI. If we don’t deliberately maintain the apprenticeship structures inherent to the software engineering practice of yesteryear (such as internships, rotations, fix-its, brown-bag presentations, etc.), we get short term efficiency while the entire long-term career pipeline collapses. We need to ask now how do we get new talent up to speed rather than some years down the line.

Otherwise, we’d get into that state that Dune tried to warn us about: Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them. Will we find the Golden Path now or are we waiting for a visionary?