Visual Creation Is Now a Conversation
There’s been a lot of noise around the latest OpenAI update—some even calling it the end of Photoshop. (Not quite.) But underneath the headlines, there’s a real and meaningful shift worth paying attention to.
As someone who works with generative AI tools daily—both in creative direction and practical workflows—I’ve seen how far ChatGPT has come with each release, I’ve seen how quickly ChatGPT has evolved. This release doesn’t stand out because of the hype, but because of what it quietly solves: Context, Consistency, and the ability to iterate—fast—all in one place.
You can now generate images directly in ChatGPT, using the new GPT-4o model. No DALL-E, no switching tools. Just type, adjust, and keep going. The model remembers your context, making the whole process feel more fluid—less like prompting, more like collaborating.
A few takeaways from my own tests:
– When uploading your own photo and asking for a style transformation—like a comic version—you need to describe the visual style in detail. Generic terms like “Marvel” won’t work; the model will block it. But if you break it down (e.g. bold outlines, saturated colors, dramatic lighting), it gets surprisingly close.
– I experimented with a simple object on my desk: a pink Victorinox peeler packaging. I asked for a mock landing page with nine color variations, each placed in front of Swiss mountain scenery, paired with an app-based shop layout. The results were better than expected—especially the product placement and visual variations. Logos still come out messy, but the fact that I got this level of specificity through text alone is something Midjourney hasn’t managed yet.
– I also tried generating images based on my own photo, asking the model to change clothing and accessories. It handled facial features quite well, though the resemblance isn’t 100%—more like a close cousin than a twin.
Overall, the image quality is good, especially for realistic lighting, product mockups, or quick storyboards. Where it still struggles: fine details (logo etc), design precision, or more complex scenes with layered meanings.
But what’s most striking isn’t the output—it’s the experience.
There’s a new kind of ease to visual ideation, where you’re not constantly starting over or re-explaining your intent. It feels more like a creative dialogue than a series of disconnected prompts.
It’s not replacing Photoshop. Yet. But it’s pointing to a future where creative work becomes more conversational, iterative, and fast.
For anyone working in strategy, design, or concept development:
This is worth exploring, not because it’s flawless, but because it’s changing how we work.
And if I had to bet on the future of design? It’s looking a lot like… typing.
