ChatGPT-5 Is Here, But Let’s Not Distracted by the Shiny New Thing

Just because because the hammer is fancy, doesn’t mean it’s better

Disclosure: I created a prompt do be a co-writer for me. It’s trained on my style, my quirks, eight years of blog posts, and the entirety of the last book I wrote (435 pages). I asked it, as a test, to write something for me about GPT-5. The first draft was really good, but I have edited significantly it to be “me.” This was a first try and I think I’ll keep using the prompt for drafts in the future. At least for now.

Admittedly, I haven’t played around with GPT-5 much. I’ll throw some things at it—maybe my new co-writing prompt—to compare it against Gemini 2.5 Pro (where I do most of my work), but this is what I’m really thinking:

Who cares.

This isn’t being dismissive, it’s pragmatic. We can’t upend all our workflows every time the new shiny AI thing comes out. Test? Absolutely. If there is a new feature you’ve been waiting for that your tool doesn’t have? Cool, check it out.

But I spend a good chunk of my day coaxing models into forming cogent sentences, doing research, and coding—and getting caught up by the latest and greatest model is a productivity trap and killer. It’s like being a carpenter who obsesses over the newest hammer with ergonomic, carbon fiber handle while forgetting that the hammer with its hickory handle still pounds nails really damn well.

The Real Conversation Isn’t About “Which Model is Best” It’s About What Gets The Job Done For You

I’ve seen a lot of smart people echoing this sentiment lately. consistently beats the drum that the right way to use AI isn’t about picking one model over another, but about building systems and processes that use the best tool for the job at any given moment. wrote that the true power of these tools emerges when you understand their capabilities and limitations, not just which one has the highest benchmark score of the week.

Jumping from one shiny new model to another shiny new model just wastes your time. You’re spending time learning the quirks of a new system instead of building things with the systems you already know. It’s the digital equivalent of constantly switching hammers instead of just swinging the one you have until it doesn’t cut it anymore.

Use the model that works for you now, while still keeping in mind that tying your entire workflow and AI strategy to a single model—whether it’s from OpenAI, Google, Anthropic, or anyone else—is just as problematic. It’s a fine line between getting to know and investing time working with a model to get the most out of it and locking yourself into a workflow that will only ever work on a specific system. You need to balance investing in learning one system while not chaining yourself to it.

A Pragmatic Approach to Navigating the AI Hype Cycle

So, what’s a marketer who isn’t part of the AI nerd herd supposed to do? The answer is easy: focus on your process, not the tool. Yes, learn the quirks of the tool you like most, but focus on “great prompts” not “a great prompt for…” Here are a few tips to stave off AI FOMO and FUD:

• Keep Your Prompts Agnostic. While I spend most of my time with Google’s Gemini and build my custom Gems (custom GPTs) there, I make sure the core prompt works across different models. I follow a set process of using a meta prompt to help me write a solid, useful prompt. I follow the same process for Gemini and a client’s Copilot instance and it always works. The prompts function and do what I’m trying to get done. Maybe I need to refine it a bit, maybe I load files a little differently, but a good prompts should be a good prompt no matter what. A well-structured prompt that clearly defines the context, task, audience, and output will get you 95% of the way there, regardless of whether you’re using ChatGPT, Claude, or Gemini. The underlying principles of good prompting don’t change just because one model got “smarter.”
• Test New Models, but Don’t Abandon Your Workflow. When a new model like ChatGPT-5 comes out, absolutely kick the tires. See what it can do. Run some of your standard, challenging prompts through it and compare the results. Does it handle nuance better? Is it faster? Does it generate cleaner code? Great. But don’t immediately dismantle a workflow that works and delivers results just because there’s a new hammer on the market. Incrementally integrate the new tool where it offers a clear advantage, rather than betting the farm on it overnight.
• Embrace “What If…?” The single greatest gift AI gives us is time. Tasks that used to take hours—like creating a dozen variations of ad copy, doing a massive content audit, deep research, or summarizing a dense report—can be done in minutes. Use that reclaimed time to explore. This is where the real transformation happens. As Christopher Penn says, we’re moving from optimization to transformation (and you really should check out his book). AI tools like Google’s Opal and custom AI agents aren’t just about doing the old things faster; they’re about enabling us to do entirely new things. Take that extra hour and ask, “I wonder if I could…” Then go find out. Vibe code a little app. Brainstorm a completely new marketing angle. Throw a crazy idea at your favorite model—or a new one—and see what comes back. That’s where the magic is.

ChatGPT-5 is fantastic. It’s a testament to the incredible pace of innovation in this field—and also shows where/when we’re hitting computational and algorithmic walls—but it’s just one more powerful tool in an ever-expanding toolbox. The real winners won’t be the ones who jump from tool to tool; they’ll be the master craftspeople who know how to build amazing things not matter what hammer they’re handed.

Now, I need to dig out my hammer to hang some pictures—a short handled one I use especially for pictures and close quarters because my regular hammer lets me swing too hard.