AI Makes Tech Less Terrifying For “Word People”

I can’t lie; as a freelance writer living through this explosion of AI tools, I finally feel…seen.

After decades of software revolving around complex code, we're finally seeing tools that feel like they’re built for people who think in stories, phrases and sentences.

But for all that’s been written (and fear-mongered) about AI, one important aspect goes under-discussed: Natural Language Interfaces (NLIs).

NLIs promise a less bitter-sweet relationship with digital tools, where the compromise doesn't feel quite so hefty. Suddenly, I can lean on my natural language tendency without needing to constantly fit into the mold of highly complex Graphical User Interfaces (GUIs). They're making tech less of a minefield and more of a playground for us humanities-inclined folks.

I work in tech, but I identify more as a "word person", which is a bit of an oddity. In school, I was an English major and kept a healthy distance from any math classes.

But I love the internet too. It's my second home. And ironically, I run a business writing sales copy for B2B SaaS companies, which means dealing with a tech stack of business tools. And honestly, their Graphical User Interfaces (GUIs) are clear as mud to me.

Are they misunderstood by me? Definitely. Am I throwing away cash on underutilized apps? Quite possibly. Do I want to spend a whole afternoon scouring help documentation or having a call with customer support to figure it all out? Not a chance.

As Steven Johnson posits in his book Where Good Ideas Come From the concept of the "adjacent possible," each new innovation unlocks a door to a previously unimagined room of possibilities. Think of it as a "Sealed Garden" – the technological garden brimming with opportunities was once enclosed behind a tall, locked gate, the key being technical proficiency. But NLIs have replaced the lock with one that uses a more commonplace key: natural language.

In a piece from the Harvard Business Review titled The Power of Natural Language Processing, the author pulls back the curtain on how NLP tools are revolutionizing his research work. They're doing more than just aiding. They're reshaping tasks like summarizing, data labeling, brainstorming, and even literature reviews.

It's still early days, but it's already clear that the potential is enormous.

Did natural language abilities just become a “technical skill set”?

The ability to express ideas clearly, to write effectively, and to think critically – these are now becoming keys to unlocking the fuller potential of the digital world. As evidenced by tools like ChatGPT and Adept.ai, these skills are being rewarded more than ever.

Recall how Netscape Navigator took the internet by storm in the mid-90s. Why? Because it made the tangled web of internet protocols way less intimidating for everyday people. For the first time, you didn't need a degree in computer science to surf the web. That was a big deal, and Netscape quickly took a big chunk of the market.

As Jason McCaffee discusses in his article, The Shape of Interfaces to Come, we've seen a significant evolution in the way we interact with technology. From command-line interfaces (CLI) to graphical user interfaces (GUI), and now to natural language interfaces (NLI), each step has made technology more accessible.

Fast-forward to today, and I reckon we're on the brink of a similar shakeup, this time with AI tools.

Beware of the reductive nature of AI models

Models like GPT-4 operate on a simple principle—they draw from the inputs provided to generate outputs. Their effectiveness is tied to the quality of the inputs they receive, sometimes even slightly less.

Imagine you're navigating the vast expanse of the internet in search of information. The information you're seeking serves as the input that the language model remixes and refines. But what happens when our focus shifts to generating new, original ideas?

Language models require a foundational idea on which they can build, remix, and enhance. By providing that centerpiece, we enable the model to contribute in ways that are scalable, surpassing the limitations of our human brain. It adds layers to your initial idea, building around it, enriching it, and combining it with other concepts.

The future of human-computer interaction is conversational, and that’s good for writers

For writers, AI has its fair share of limitations — hallucinations, biases that echo its training data, and a plethora of bugs inherent to an early-stage tech. Yet, despite these hiccups, I think there's a lot to look forward to. Personally, I'm a fan.

Stumble upon your own set of discoveries that highlight what it does well, and maybe even what you struggle with. Acknowledge that AI is a different kind of "brain" — a complex one, yes, but not nearly as intricate or nuanced as the one sitting in your skull.

And, crucially, understand that AI can scale its thinking far beyond what any human can do. It sifts through vast amounts of data, pinpointing the needles in the digital haystack. Take these insights, these little nuggets of gold, and use them to breathe even more life, more humanity, into your work.