On Friday morning, I opened my laptop to begin work. I’d left ChatGPT open the day before when I finished work and hibernated my device. As the screen lit up, Chrome restored my last tabs. But the ChatGPT on my screen wasn’t the same one I’d left yesterday. At the top of the home page, it read: Introducing GPT-5.
I looked again to make sure I wasn’t mistaken. “If OpenAI had launched a new model of ChatGPT, then surely someone must have covered it on the Innovation Village blog,” I thought. So I went to the website to confirm, and it was true that OpenAI had launched GPT-5.
My first prompt to ChatGPT-5 wasn’t about work. I just wanted to see how it would respond to something light and slightly challenging. The reply was sharper and more engaging than I expected, and it was immediately clear this new model could help me work more efficiently.
My Initial Experience
The first thing I noticed was how smooth the interaction felt. Though the responses were fast, it felt like talking to someone who already understands how I think. I entered a rough paragraph, and instead of simply correcting the text, GPT-5 asked me questions to clarify my intent, offered different perspectives, and highlighted a fact I should double-check. That moment made me realise this wasn’t just an update. It was a smarter, more intuitive tool that genuinely enhances my workflow.
Why This Feels Different to Me
I’m not a software engineer, unless you count the times I yell at my laptop when my Wi-Fi drops. I don’t use ChatGPT for debugging or code reviews. My work revolves around the use of words, ideas and planning:
- Writing: articles, essays, and personal reflections.
- Editing: turning rough drafts into clear and polished writing pieces.
- Developing Ideas: breaking down big or complex information into simple ideas.
- Research: going beyond simple searches to connect scattered facts into useful information.
- Staying Updated: learning new topics, following trends, and refreshing old knowledge.
With GPT-5, all these tasks become faster, more precise, and efficient.
What’s New?
- Nuanced Context Memory: It now remembers and reuses small but important details from earlier in the chat, the kind GPT-4o often forgot.
- Better Self-Critique: When I tell it to review its work, it actively spots tone slips, wordy sentences, and unclear facts, without me having to nudge it on every point.
- Multi-Step Reasoning: Research now comes with the “why” and “how.” Instead of just dumping findings, it walks me through the logic that connects each source.
- Creative Range: Storytelling and voice-shaping have levelled up. It can match my casual-but-precise style all the way through, without sliding into flat, generic AI language.
The Real Upgrade
The magic is not in a single feature but in how the technology easily becomes part of my process. With GPT-4o, I felt like I was operating a tool. With GPT-5, it feels like working alongside something that gets how I think and still manages to surprise me.
That’s why, as I finish this review, I’m aware that half the difference is the model itself, and half is how it fits seamlessly into my process.
