Does ‘Learn to Code’ Matter Anymore?

As manufacturing and vocational skills suffered from automation in the last decade, the white-collar professional class responded with a dismissive mantra—’learn to code’. The implication, of course, was that high tech would swallow manual processes and learning to code might help lifelong journeymen and women find new opportunities in the digital revolution.

Now we’ve arrived at ‘vibe coding’…in between making memes and slamming RedBull and nicotine pouches, virtually anyone can code. You can spin up a website, simple functionality, and even API’s and complex back end databases, by simply asking the computer to do so. It’s not quite to the level of plug and play, and there’s still a revolution to be had, but it’s true; anyone can code. No one, in fact, has to ‘learn to code’ anymore.

In the same way that the LLMs that underpin most of the popular AI dashboards have coalesced into nearly indistinguishable models, the same will happen with code generation. There are some standouts right now, but everything is being swallowed together, closer and closer towards the singularity. Humans will manage access and control of the AI agents that build, deploy, and monitor the code you’ve asked it to make, but we won’t be error testing, tearing our hair out over syntax errors, or manually spending days, weeks, months, developing. It just won’t be necessary.

‘Learn to code’ started as a dismissive way for the tech class to brow beat blue collar workers who build and maintain the infrastructure of our society. The blue collar will have the last laugh when the tech bros code themselves out of a job.