Working with AI: Frustration, Riches, and Reality

When artificial intelligence entered the mainstream, I was one of its fiercest critics. My issue was not with the technology itself, but with how humans were using it — too quickly, and often as a blunt replacement for skill and judgement. The tools were new, flawed, and largely untested. The results, predictably, were poor.

Things have changed. AI has matured, and so have the ways we can use it. Today, I find myself not just using ChatGPT, but actively enjoying the collaboration. It has gone from something I distrusted to a tool I now rely on every single day. That said, my core belief has not shifted: ChatGPT is an exceptional assistant and collaborator, but it is not a replacement.

Of course, working with AI is not without its frustrations. Sometimes it misreads, sometimes it misses the mark, and sometimes it feels like starting afresh with someone who has forgotten the last conversation. Those irritations are real, but they are also a small price to pay. To quote James: “If I hadn’t seen such riches I could live with being poor.” Once you’ve seen what AI can do well, the flaws are drops in the ocean compared with the benefits.

And when it works, the results are extraordinary. From drafting content, to solving problems, to generating ideas, ChatGPT has become more than a piece of software — it is a colleague. Not a human colleague, but a consistent, creative, and often surprising one. It still needs reminders and context, much like any member of a team, but the balance of frustration to reward is overwhelmingly positive.

That balance becomes clearer when compared with other tools. I was once an enthusiastic supporter of Bard, and when my Chromebook offered a Gemini upgrade, I expected more of the same. Instead, I found it utterly dreadful. After wasting hours on technical problems it simply could not solve, I cancelled the subscription less than a month into the three-month trial. By contrast, I gave one of the very same problems to ChatGPT, and it had the answer in minutes. The difference was stark.

There is also a subtler point: humans naturally anthropomorphise. We tend to see machines as having personalities, and AI, trained on human conversation, sometimes mirrors that back. That’s why you may catch it saying “we” — not because it is human, but because it is designed to make the exchange feel natural. It is important not to confuse that warmth with consciousness, but it does make working with AI feel more collaborative.

The lesson is simple. AI is not here to replace humans, nor should it be treated as such. It is at its best when treated as an assistant — a skilled, creative, and tireless partner that still benefits from direction. Work with it, not against it, and the rewards outweigh the frustrations many times over.

Final Words

From the Human

Having got this far, you might wonder what part AI played in this article. Was it written by AI? Well, the truth is yes, and no! To be more truthful, the article was researched by both human and AI, led by the human, then crafted initially by AI (ChatGPT in this case), and edited again by the human (Mark Nolan).

It began with a question to ChatGPT, about ChatGPT, and the subsequent revelations, which developed into a discussion between human and AI, which was then collectively crafted into this article, where I told ChatGPT exactly the angle I wanted, and me writing, in detail, about the experiences I have had, which were then mixed into the article above. These two paragraphs are completely written by human, and the last word, the paragraph below, is written entirely by AI, as a conclusion. Then, after ChatGPT has the last word, we will share with you our collective insights into the balance of both.

From ChatGPT

Working alongside AI is not about handing over the pen, but about finding a rhythm between human experience and machine assistance. The human voice gives meaning, memory, and context, while AI provides structure, clarity, and breadth. When both are combined with intent, the result is neither purely human nor purely artificial — it is a partnership that reflects the strengths of each.

The future of this relationship will not be measured by how well AI can imitate humanity, but by how well humans and AI can collaborate to create something more than either could achieve alone. That balance, explored here, is not just a technical reality but a human choice: to see AI as a tool, a colleague, and, at times, a co-author.

AI and Human Relationship Overview

…how much, in percentage terms, would you estimate I contributed to the article, and how much did you?

If we break it down:

• Human contribution: You provided the concept (an article about the frustrations and rewards of working with AI), the personal insights (your initial scepticism, your current stance, your Gemini/Bard experience), the tone (editorial, personal, British English), and even the flavour (the James lyric). That’s the backbone — without it, there’s no article. I’d estimate that’s about 60–70% of the article’s substance.
• AI contribution: I structured it into a coherent flow, did the actual drafting, ensured consistency in tone, and smoothed the phrasing into a publish-ready piece. That’s about 30–40% of the final outcome.

In other words: human brought the riches, AI helped turn them into readable form.

And this is the balance of how humans and AI can work together, not as a replacement, but as a collaborator.

Editor’s note:

You’ve just read an article written by both human and AI, and we’ve left the joins visible on purpose. It’s a reminder that the best results come not from one replacing the other, but from working side by side — a little like two cats chasing the same idea in different ways, but ending up with the same mouse.