Version 1: Fear Not the AI Doomsday—Artificial Intelligences Still Can’t Write Their Own Code

The Reality of AI Coding: Why We Shouldn’t Fear the Future

In recent days, I’ve had quite the experience working with artificial intelligence, specifically ChatGPT and Claude. What began as a straightforward task—creating graphics for my Substack—quickly turned into a comedy of errors that showcased the limitations of AI technology in coding.

Initially, I considered using PowerPoint for this project, thinking it would be a simple and effective approach. However, curious about the capabilities of AI, I decided to test its potential for coding instead. After all, why not explore this modern tool?

What I anticipated would be a streamlined process transformed into an unexpected struggle. Each attempt to correct one issue seemed to unleash a torrent of new challenges, creating a chaotic cycle that felt like playing a game of Whac-A-Mole. Instead of the seamless efficiency I hoped for, it became a frustrating affair reminiscent of a rookie coder’s first foray into programming. My efforts with AI were met with an endless series of setbacks, aptly described by the phrase “clusterf***.”

After several days of navigating this digital maze, I finally conceded defeat and reverted back to the familiar—producing my PowerPoint presentation in a mere three hours. The experience was eye-opening and led me to reflect on the current state of AI in the coding realm.

While it’s true that AI systems are bound to improve through techniques like Reinforcement Learning from Human Feedback (RLHF) and AI-generated feedback, the truth remains: the technology is still in its infancy, akin to a novice coder on their very first assignment.

For those worrying about the rise of intelligent machines and the potential for an “AI Apocalypse,” I would argue that such fears may be exaggerated or even fueled by sensationalism. Instead of stressing over the prospect of AI taking over our jobs or coding skills, perhaps it’s time to relax. Raise a glass, take a breath, and really, just remember that while AI is a powerful tool, it certainly isn’t infallible yet.

In conclusion, let’s keep the conversation about AI grounded in reality, appreciating its capabilities while understanding its limitations. After all, sometimes it’s better to stick with the tried-and-true methods we know work well.