Use This GPT to Make DALL-E More Realistic Than Midjourney

A simple GPT is all you need to make images in DALL-E that are even more realistic than MidJourney can muster. The Photorealistic GPT will transform your images, and might help you cancel the Midjourney subscription. 

Try the Photorealistic GPT from GeniGPT here.  

And check out a few of my efforts here:  

Prompt: Happy successful family enjoying food by the pool. The father is a mid 30s businessman, the mother is blonde and the kids are 9 and 7 years old.

Prompt: A black Instagram influencer model walks down the street in Cuba.

Prompt: A male film lead star type looks down over a broken city. There’s a grimy and hyper real aesthetic.

Prompt: Latina influencer, about 22 years old, and she’s in the gym. Shes beautiful and slim.

Prompt: Close up shot of an Indian girl, 19 years old, in full make up and dress for a religious ceremony.

Prompt: Japanese businessman with a city skyline and volcano behind him. In his hand is a traditional Japanese sword.

Prompt: A homeless man with a sign asking for change in the streets of New York.

Prompt: A Bugatti Chiron in the street.

Prompt: A racing bike in the Isle of Man TT races.

Prompt: A heavyweight boxing match, one fighter is taking a punch.

Suddenly DALL-E is Good Enough

You know as well as I do that subscriptions mount up, and this makes ChatGPT ‘good enough’. So anytime you can save money AND get the same quality illustrations. It has to be a good thing.  

I mean our collective jaws hit the floor when the likes of DAll-E, hit the market. Finally we could forget the stock images and create the perfect graphic or picture to go with any story or sales pitch. 

But it wasn’t quite like that was it? I mean text-to-image generators have been a revelation, but they also turned pretty generic, pretty fast.  

The spelling mistakes are still part of the DALL-E experience, even with the GPT, but there are ways to beat that too.

There’s also a cartoonish element to naked DALL-E that just doesn’t quite fit the corporate blog environment especially. These pictures come out looking like realistic paintings, or flat out animation, and that’s a hard sell to corporate lawyers and serious old-school companies.  

Lots of us switched it up to Midjourney and Lensgo.ai.  

They are both still superior by the way, with lots more options and solid text-to-video software coming on line. So if you need the best, you still need them. This technique really just makes ChatGPT a lot better. 

We’ll do a separate post on getting the most out of DALL-E, but this is a really quick post with a really quick solution to getting more realistic images out of ChatGPT and smoothing out your workflow.