Imagine creating beautiful visuals just by describing them in plain language – no complicated tools, no waiting for graphic designers to send you drafts back and forth, and no complaining. It’s sort of the ultimate dream, and despite the tremendous advances in AI, it’s always seemed so far out of reach.

However, OpenAI’s latest rendering model (4o) is changing that. It’s arguably brought us closer to the promised land, or even right into it. This new model is breathtaking. It seems to do everything right. It can create realistic images, doesn’t distort shapes, understands context, and has turned everyone into a designer overnight.

Let’s start from the beginning and ask the most obvious question. Why does it work so well, and why can’t other models replicate it? If you want a simple explanation, GPT-4o is an “omnimodal” model, meaning it understands and connects different types of data, including text, images, and even audio and video. It also understands prompts better because it’s trained to grasp the meaning behind words and images, rather than strictly following the exact words given. Other models often follow prompts word for word, so you’ll need to provide more detail to get the right result. But GPT-4o can guess what you mean even if you don’t say it all outright, making the images more natural fits your intentions.

They’re also different from diffusion models. These models work by starting with random noise and slowly transforming it into an image based on the prompt. Every time you want to change something—even a small part—you’ll usually have to recreate the entire image or use special techniques to edit it. They’re great at creating high-quality images, but they’re not very interactive or smart at making incremental changes.

The new 4o image generator is a token-based model that treats images more like language. It “sees” the image as a series of tokens (like words) and can understand and update specific parts of it, like editing a sentence. So if you say “add a tree” or “move a chess piece,” the model can change just that part of the image without touching everything else.

And because he already knows so much about the world, he’s really good at what he does. This will undoubtedly change a few things. Marketers and businesses will now have a powerful tool to create and iterate visuals in real time without needing advanced design skills. For example, they can dynamically create product images in different colors and styles, tailored to individual users. They can run extensive A/B tests, or customize entire ad campaigns by explaining any updates in plain language. Ecommerce platforms can personalize product images to the customer, while product teams can quickly prototype UI and UX designs by instantly turning rough ideas into visual mockups. Creators, meanwhile, will benefit from the ability to instantly adapt or improve their thumbnails.

There’s something for everyone. OpenAI included. This will definitely help them attract a lot of new users, and as more users engage with the platform, OpenAI will collect more data to continually improve its models. This could be a virtuous cycle and give the company a big advantage as the AI ​​race gains momentum.

However, there’s one problem we haven’t discussed yet. What does this mean for designers and graphic artists? If everyone can create engaging visuals with a simple command, will their jobs be over? Well, people often exaggerate the catastrophe in these matters.

People in the 19th century believed that mechanized looms would eliminate textile jobs, but the industry expanded and created new factory-based roles instead. The introduction of ATMs was supposed to make bank tellers redundant, but their numbers grew as banks redeployed their employees to customer service and sales roles. Personal computers, initially seen as potential job killers, have given rise to entirely new industries in IT and software development. Even e-commerce, once feared as the end of traditional retail, has evolved into an integrated model where traditional brick-and-mortar stores and online platforms work together.

People adapt. They do. And guess what? Designers will embrace these new developments and incorporate them into their workflows. The only question is: Who is smart enough to do it quickly? We’ll have to wait and see.