Reasons To Learn Programming

You know this guy. He got in before tech required a CS degree and a LeetCode black belt, rode the dotcom wave, and now makes six figures while asking "Claude..." in every meeting like he's summoning a genie. Hasn't touched code since dial-up was fast, but absolutely convinced he could still outcode the entire dev team if he "had the time." Meanwhile he's dropping 120k on a smartwatch and would literally risk it all for Claude Anthropic's API. The shoes that have "been at the same company for years" really sell it—comfortable, broken in, going nowhere. And that weird hobby? Probably collecting vintage keyboards or explaining blockchain to his neighbors. The best part? He genuinely believes his IQ is 140+ because he solved IT problems in an era when turning it off and on again was considered wizardry.

When your AI co-pilot decides to inject inappropriate jokes into your production code! The meme shows Elixir/Phoenix code with a logger statement containing "Dose nuts fit in your mouth?" - that classic middle-school joke now immortalized in your codebase. Imagine deploying this to production and then having to explain to your manager why your app is making "deez nuts" jokes in the logs. That PR review is going to be... interesting. 💀

Nothing says "we love our users" like dropping a free AI feature that immediately sets fire to your data center. Those poor GPUs, running at 110°C, fans screaming like they're auditioning for a metal band. Meanwhile, DevOps is frantically calculating the electricity bill while the marketing team high-fives over user engagement metrics. The best part? The feature probably could've been implemented with a simple if-statement, but hey—gotta justify those VC millions somehow!

The job market has evolved from "write code" to "convince the AI overlords you're human enough to review their hallucinations." Turns out companies realized they don't need engineers anymore—they just need someone to babysit ChatGPT and tell it when it's being confidently wrong. Senior Prompt Reviewer: must have 10 years of experience with GPT-5, expert at saying "no that's not how arrays work" to a language model, and willing to work for equity in a startup that's definitely not just three LLMs in a trench coat.

ChatGPT just solved a simple algebra problem by literally writing code in natural language. Instead of setting up basic equations (sister's age = 3 when you were 6, age difference = 3, so sister = 70 - 3 = 67), it decided to... evaluate mathematical expressions as string templates? The <<6/2=3>> and <<3+70=73>> syntax looks like some cursed templating engine that escaped from a PHP nightmare. The best part? It got the answer completely wrong. The sister should be 67, not 73. But hey, at least it showed its work using a syntax that doesn't exist in any programming language. Our jobs are indeed safe when AI thinks inline computation tags are a valid problem-solving approach. This is what happens when your training data includes too much Jinja2 templates and not enough elementary school math.

So we're automating ourselves out of jobs, only to realize we need those jobs back to handle the automation we built. This is basically the tech industry's version of "what came first, the chicken or the egg?" except both are getting replaced by AI and neither of us can afford rent. The beautiful irony here is that we spent decades convincing truck drivers that robots would take their jobs, and now those same truck drivers might end up being the ones maintaining our legacy codebases while we're out there delivering Amazon packages in semi-trucks. The simulation is glitching. Plot twist: by 2036, the AI will automate truck driving again, and we'll all become coders once more. It's the circle of tech life, and it moves us all. 🦁