Neither Ghibli nor Teleñecos, the new chatgpt image generator is impressive and you are wasting time

It is said that artificial intelligence will replace human beings, but perhaps the one that is being most affected with the current situation is the same tool, due to the massive image phenomenon based on classic anime style that is trend.

In social networks such as Facebook and Instagram, many users are sharing their versions based on the study style that created Chihiro’s tripbut perhaps people are overlooking the true potential of the GPT-4 model, because it is an advance that goes beyond, it is even superior to other chatbots.

If you have invested in a chatgpt plus or PRO subscription, then you should take advantage of all the capabilities that this technology grants. In itself, generating images with AI is relatively simple, but you could get even more impressive results if you learn some tricks.

Improves the use of prompts

First of all, the most important thing when using Microsoft Copilot, Deepseek, Google Gemini or, in this case, Chatgpt, is the structure of the prompts. While things like attach a photo and tell him to make it «Ghibli style» Or any other artistic approach is not the most optimal.

That action is too basic, then the way in which you conform the image application influences the result. The best thing is that you follow a composition such as highlighting “make an image” or “convert this image”, depending on the case, and continue with the details of the subject, the actions or objects used, the scenario and the style.

For example, the image you see has been created with the prompt «Make an image of a pirate with a blue hat and a patch in the eye, that is on a raft and hold a flash in your right hand. The pirate is looking from afar an island in full sunset. That is hyperrealist style.»

If you follow these parameters in the same order, With practice you will create better designs with the help of Sam Altman’s artificial intelligence.

Try other designs beyond Ghibli

Ghibli undoubtedly is one of the most great designs you can get with AIbut it is not the only one in which you should focus with the same photo or with the creation of a from scratch, you can have incredible results that will surely become your favorite designs.

There are a wide variety of styles and Everything is a matter of experiencing with them with the recommended indications In the previous Council. If you don’t want to get away from Hayao Miyazak’s works, then you can opt for alternatives such as Shonen Anime, Black and White Manga, Chibi, Kawai or Moe.

However, there are no limits, you can Use other artistic forms in chatgptsuch as Voxel, oil painting of different designs, abstract, pixel art, pixar art and many more. If you have good precision with the phrases in the PROMPT, it will not disappoint you.

Edit the images generated by the AI

Didn’t you like the result you have given you? Then try edit the image from the same chatbot. That is, if you requested a design and that has been good, but there is something specific that does not convince you, you can change it by doing another continuous prompt.

An example is that you have asked to make «an old man sitting on the shore of a beach», but There are things you would like to change. If so, it is possible to ask you to add some «sunglasses» and have a more serious outfit or that «green aliens are accompanying it.»

In turn, you can also copy or edit the prompt if something has not done well Or you would like to change the style to a completely different style. This is something that definitely opens endless possibilities when designing, since little by little you have the ease of adding or removing things until you reach what you need.

Add text to the photos

One of the biggest steps that GPT-4 has given is that now There is the ability to add text to the images in a precise wayand even products or images based.

This was unfeasible because it had many limitations and I used to create hallucinations with all outdated letters or languages ​​were linked and the phrases were incomplete. Now, even an image of a full wine glass can be generated with this technology and that should also take advantage of it.

In photos with posters, electronic devices, t -shirts or in any other way, you just have to go to the phrase “that has the following Spanish text: ‘text’” and determine where the image should go. The answer is good, although Not if it’s too extensive phrases.

Change the appearance relationship

The OpenAI tool generates images in relation 1: 1 (square) appearance, but it does not have to be strictly like that. At the end of each prompt you can add this information To adapt to the format you need.

There are many types of standardized appearance relationships, such as 16: 9 or 22: 9 for a computer wallpaper, 9:16 for mobile or 4: 5 for Instagram verticals and other social networks. If details in your application, You will save the process of having to cut it or have to expand it with another AI tool.

As you can appreciate, the world of generating images with chatgpt is extensive, but with these tips, the way of making content will change completely and significantly improve your experience.

Know How we work in NoticiasVE.

Tags: Photography, advice, artificial intelligence, design