AI and Accountants: Why We're Still Arguing About Whether ChatGPT Should Write Emails

Last night, in a moment of weakness (and insomnia), I did the unthinkable: I opened a Twitter thread. The tweet that lured me in was harmless enough: "Why hire an accountant when there's AI?" But Twitter threads, especially about tech, are like opening a cupboard you know is empty—there's nothing in there except disappointment. And boy, was I disappointed.

The replies were… well, exactly what I expected. "AI can't replace accountants because it's not human," and "AI can only do bookkeeping, not actual financial analysis." Loud. Confident. Wrong. Reading through the comments, I felt my frustration rise. It wasn't just the misinformation—it was the sheer audacity of people making these claims without even a basic understanding of what AI is capable of. And that's when it hit me: this is exactly the problem with how society understands AI. On one side, you have people like me—deep in the trenches, witnessing the wonders of this technology. And on the other side, you have the general public, still debating whether letting ChatGPT help you draft an email makes you "lazy."

Why It Made My Blood Boil

Let me explain why this thread hit a nerve. I'm not just some casual observer. I've spent years studying computer science, working with machine learning, and supporting AI services at Amazon. I've seen AI do things that would make your jaw drop—tools like DeepSeek R1, a Chinese competitor to OpenAI, which is pushing the boundaries of NLP and data processing. I've worked with Claude, a conversational AI so advanced it feels less like software and more like brainstorming with a genius colleague. And then there's DALL·E, an AI capable of generating breathtaking images from just a snippet of text, revolutionizing creativity in ways we couldn't have imagined a decade ago.

The gap is enormous. On one side, you have people like me—AI engineers, researchers, and tech insiders who've seen the absolute wonders these tools can achieve. We've worked with systems powered by transformer models, the architecture behind breakthroughs like ChatGPT and Claude, that revolutionized natural language processing. And on the other side, you have the general public—people whose understanding of AI begins and ends with ChatGPT. They don't know about DeepSeek R1, transformers, or the incredible advancements coming out of labs like Anthropic. They're stuck thinking of AI as "that thing that helps with emails."

And it's not just ignorance—it's confidence. Confidently proclaiming AI's limitations while remaining oblivious to the fact that these limitations have already been shattered. That's what stings.

The Knowledge Gap: Why It's So Massive

The gap isn't just about knowledge—it's about the pace of progress. AI is evolving so fast that even we in the tech field struggle to keep up. But for the general public, it's like trying to binge a complex sci-fi series when you've only seen the trailer. Here's why this divide is so vast:

1. Exposure issues: Most people have never come across tools like Claude, DeepSeek R1, or DALL·E. They don't know about revolutionary concepts like transformers or generative adversarial networks (GANs). For them, AI is synonymous with ChatGPT—end of story.
2. Media narratives: The media doesn't help. It's either "AI is stealing your job!" or "AI can't even write a decent essay!" There's little focus on how AI is fundamentally reshaping industries.
3. Fear and skepticism: People fear what they don't understand. AI feels abstract, distant, and, frankly, intimidating. Instead of engaging with it, they dismiss it.
4. Educational gaps: In South Africa—and globally—tech education lags far behind AI's advancements. Most people don't have access to the resources needed to even begin understanding the power of these tools.

What AI Can Actually Do (And Why It's Mind-Blowing)

Let me give you a glimpse into what AI is really capable of:

• DeepSeek R1: This Chinese OpenAI competitor is making waves with its ability to process complex datasets, deliver unmatched natural language processing capabilities, and rival even the best Western models.
• Claude by Anthropic: Claude is like the genius sibling of ChatGPT. It's capable of nuanced conversations, advanced problem-solving, and even creative tasks. It's not just an assistant—it's a collaborator.
• DALL·E: Imagine describing a scene—"a cat wearing a spacesuit, sitting on the moon"—and seeing it come to life in a stunningly realistic image. That's the power of DALL·E, a tool that's transforming creativity and redefining what's possible with AI.
• Cursor: If you've ever written code, you know how painstaking it can be. Cursor makes that process effortless, generating, debugging, and optimizing code with uncanny accuracy.
• Transformers: The backbone of modern AI, transformers power everything from text generation to image recognition, making tools like ChatGPT, Claude, and DeepSeek R1 possible.
• Affective computing: AI isn't just logical—it's emotional. Tools powered by affective computing can recognize and respond to human emotions, enabling breakthroughs in education, therapy, and customer service.

And these are just a few examples. The truth is, AI is already doing things that would have seemed impossible a decade ago. But the general public? They're still debating whether using ChatGPT makes you lazy.

The Responsibility of Builders Like Me

As someone working in this space, this gap isn't just frustrating—it's terrifying. As a startup founder, I'm building AI-backed solutions for a market that doesn't fully understand their value. How do I create tools for people who still think AI is limited to writing emails? How do I sell innovation when my target audience is skeptical, uninformed, or outright dismissive?

It's not just a business challenge—it's a societal one. If we don't close this gap, we risk leaving people behind. And that's a problem not just for individuals but for entire industries and economies. The future is being built now. If people don't catch up, they'll miss out—and so will we.

Bridging the Gap: Where Do We Start?

Here's how we can start addressing this divide:

1. Make AI accessible: Tools like Claude, DeepSeek R1, and DALL·E need to be more than just buzzwords for tech insiders. They need to be approachable and understandable for the average person.
2. Educate, educate, educate: We need to demystify AI. That means breaking down complex concepts into simple terms, offering workshops, and creating content that engages the public.
3. Shift the narrative: AI isn't here to replace humans—it's here to augment our abilities. This is the story we need to tell.
4. Lead with empathy: It's easy to get frustrated with misinformation, but the solution lies in patience, not judgment. We need to meet people where they are and bring them along for the ride.

Final Thoughts

That Twitter thread wasn't just frustrating—it was a wake-up call. It reminded me how far we've come in AI and how far we still have to go in bringing the world along with us.

The tools we're building—Claude, DeepSeek R1, DALL·E—are powerful beyond imagination. They're reshaping industries, solving problems, and redefining what's possible. But if the general public doesn't understand them, their potential is wasted.

As a startup founder, this gap is daunting. But it's also an opportunity. An opportunity to educate, innovate, and create tools that bring the future to everyone—not just those of us in the know.

So, to the people on that Twitter thread: AI isn't just writing your emails. It's building the future. And it's time we all caught up.