AI's Impact on Coding Skills: A 17% Decline?

In the grand tradition of technological advancements, AI is here to save us from the drudgery of coding—by making us worse at it. According to a recent study by Anthropic, AI-assisted coders scored 17% lower on a post-trial test than their unassisted counterparts. Think of it like switching from playing Duck Hunt on the NES to watching Twitch streams of someone else playing Duck Hunt. Sure, you get to see the ducks fall, but are you really honing your aim?

The Anthropic Study: A New Chapter in Tech Fatigue

Anthropic's randomized controlled trial didn't exactly reveal a plot twist worthy of M. Night Shyamalan. Participants using AI tools completed tasks slightly faster—about 2 minutes quicker on a 25-minute task—but scored significantly lower on a follow-up quiz. This study has been likened to discovering that eating donuts while jogging doesn't actually help your marathon time. For those interested in the detailed breakdown, the original study was published by Anthropic.

Productivity vs. Mastery: A Battle Older Than the Internet

The dilemma here isn't new. From Atari's Pong to today's Fortnite, every new technology promises to make our lives easier at some cost. The Anthropic study highlights a trade-off between productivity and skill mastery. "Obviously using a tool like this that significantly improves your leverage and the ease with which you do the work is going to make you worse at the core skill," Nick Saraev explains in his video.

One could argue that the convenience of AI parallels the rise of the microwave in the 80s. Sure, you could have a hot meal in minutes, but somewhere along the way, we lost the collective ability to make a decent chicken cordon bleu from scratch.

Horizontal vs. Vertical Growth: The Soap Factory Analogy

In the video, Saraev draws a distinction between horizontal and vertical growth—a concept as integral to business as Mulder is to Scully. Horizontal growth, facilitated by AI, allows us to produce more of what we already know. Vertical growth, however, requires innovation and the development of new skills. It's like choosing between opening more Blockbuster stores or inventing Netflix.

Saraev suggests that while AI can help us "grow horizontally" by speeding up repetitive tasks, it does little for vertical growth—our ability to innovate and develop new skills. "For example, if horizontal growth is going from one soap-producing factory to three, vertical growth is coming up with a new method that allows you to produce what would once have required a thousand factories," he notes.

The Changing Nature of Work: A Look Forward

The real takeaway here might be what this means for the future of work. As AI continues to evolve, it's poised to outperform humans in many economically valuable tasks. But as Saraev humorously points out, "How about talking to your wife? I'd recommend talking to your wife yourself, not passing that on to Clogbot."

The key, then, is not to outsource our humanity. We must choose which tasks to pass on to AI and which to keep, much like deciding whether to watch "Seinfeld" reruns or tackle a new streaming series. In a world increasingly dominated by AI, maintaining our unique human skills—like critical thinking and personal relationships—becomes vital.

So, where does this leave us? Much like a hunter-gatherer plucked from the past and placed in our digital age, we may question the value of our work. Yet, as we adapt, perhaps we'll find that the essence of our roles remains the same: to learn, to innovate, and to connect as only humans can.

By Mike Sullivan