Does The AI Race Have a Finish Line?

For a long time, ChatGPT was the tool. Everyone had a license, everyone was using it, and it felt like the obvious choice.

Then Gemini and Claude came along and put ChatGPT to a test.

Then Claude seemed to just take over in most categories. Better models, better features, Claude Code, Skills, Artifacts, Cowork before anyone else was doing it. A lot of people, myself included, quietly shifted away from OpenAI. ChatGPT 4 was still great in my opinion. ChatGPT 5 was utterly underwhelming. It seemed like OpenAI just couldn't get anything right.

And then, within just a few days, everything seemed to have changed again.

OpenAI dropped two things in quick succession: a new image generation model and ChatGPT 5.5. And I'll be honest, I wasn't expecting much.

I was wrong.

So let me show you exactly what I did, and more importantly, what you can do with it, too. I'll focus on image generation as this one truly stood out.

Here's exactly what I tested, step by step for image generation:

Step 1: Deep research first. I used ChatGPT to do a deep research pass on a topic I wanted to visualize. Not just surface-level, proper structured research that gave me real substance to work with.

Step 2: Generate the image prompt using my own CI. This is where it got interesting. I have skills set up at the system level that include my brand voice, my visual identity, the way I communicate. I had ChatGPT generate the image prompt using those specs, not a generic prompt.

Step 3: Generate the image. The first output? Already at roughly 80% of what I would expect from a finished piece. The text was accurate. The design adhered to my specs. The infographic was genuinely usable.

Of course, there were things that could have been better. But those I actually fixed with better prompts. So it was me, not the tool. I ran a few more iterations and kept note of my lessons learned. That way I was able to generate quite a number of spot on infographics.

This is a completely different experience from where image generation was even six months ago, where edits were inconsistent, typos were everywhere, and getting something on-brand felt like a battle. Google's newer models had already moved the needle significantly. But this felt like another leap entirely.

A few ideas how you can use Image 2.0 in real life:

1. Info flyer based on your website: provide the AI with your website, let it know what it should focus on, turn on Image 2.0 and let it generate the flyer (you can specify the aspect ratio, provide it with more info on your branding, favorite font sizes and colors etc.)
2. Ideas for social media graphics: share your post, required specs, your branding info and any ideas you have for the kind of graphic you want and let Image 2.0 do its magic. If you need inspiration, use the chat functionality first; that way you will have a better starting point.
3. Digital assets for your products and services: provide the AI with information about the product, let it know what kind of digital asset you are after, use it to ideate or go straight to image generation (don't forget to provide required specs, brand elements, etc.)
4. Pro Tip: At the time of writing this edition, Canva is still faster for small changes to colors and text and in many cases, the free account should do. 

My next step?

I think a skill is in order that helps me build those image generation prompts faster and more consistently. Maybe even multiple skills with different personas, to reflect the nuances in my audience. I cannot wait to get those in place 😇

Your next step?

Well it depends. If you have a ChatGPT subscription or are using a multi-LLM chatbot such as Langdock (which is what I am using), go and play with it. I am pretty sure you will find it genuinely useful. If you are using Claude or any of the other AI Chatbots out there, this is not a reason to move your entire set-up over to OpenAI. 

Far from it, actually. AI tools will keep leapfrogging. And if you leapfrog with them, you will spend your time chasing tools instead of doing the work that actually evolves your business.

Key Takeaway

The AI landscape moves fast. Someone will always be launching something new, something shinier, something that sounds like it changes everything. And sometimes, it genuinely does.

But most of the coaches and consultants I work with do not need to overhaul their entire setup every time a new model drops. What matters more is being consistent and focusing on what adds value to your daily work.

And if you want to experiment with new tools and features? Have a flexible system, one where you can test new tools without dismantling what already works.

Stay curious. Keep building. Test things. But don't panic-switch.

Til next time

Elena