The popularity of OpenAI’s large pre-trained language model ChatGPT has triggered widespread discussion and fierce competition in artificial intelligence around the world.
Local time on April 17, Twitter CEO Elon Musk accepted an interview with Fox News and responded positively to recent questions about his startup “X.AI”: “I’m going to start something which I call ‘TruthGPT’, or a maximum truth-seeking AI that tries to understand the nature of the universe.”
He said TruthGPT “might be the best path to safety” that would be “unlikely to annihilate humans”. According to media reports, Musk registered a new AI company called X.AI Corp. in Las Vegas last month. Currently, Musk is the only director of the company. X.AI Corp. has authorized the sale of 1 billion shares. In addition, he poached several engineers from Google’s AI department, targeting OpenAI.
However, before that, Musk had repeatedly revealed his worries and distrust about AI in public. In March this year, he also co-signed an open letter with thousands of industry executives and experts calling for a halt to the development of AI systems more powerful than GPT-4 for at least six months. This appeal has also received a lot of attention and responses around the world. Earlier, in December last year, he tweeted that OpenAI’s ChatGPT was “scary good, We are not far from dangerously strong AI”.
In the AIGC era, AI should not exist as a competitor to human society, or AI was created to serve humanity, not to cause anxiety and panic. Since 2022, AI painting, AI voice and other AI content creation have set off a wave on the Internet, especially the most popular AI painting. Midjourney, Stable Diffusion, NovelAI and other AI painting tools have emerged, and countless “fake” AI paintings have shaken the entire Internet content creation circle.
However, these popular products still have many details can be improved and room for technical iteration and upgrade. For example, two commonly used AI painting tools, Midjourney and Stable Diffusion.
In early April, Midjourney released Niji-journey V5, a version for anime. Niji V5 supports stylized parameter settings. Compared with Niji V4, Niji V5’s paintings are more realistic. In Midjourney V4 version Niji cannot use stylize commands, while Midjourney V5 can use stylize commands and refer to different artists’ design styles. However, according to user feedback, obvious flaws can be seen in some character images.
At the same time, Stability AI recently announced that it is starting public testing of Stable Diffusion XL under development. Emad Mostaque, founder and CEO of Stability AI, said the model is still in training and will be open source once the parameters are stable.
In the past, Stable Diffusion has always had many problems in generating human anatomical structures. “Missing arms and legs” is a common problem. For example, SD-v1.5 often generates distorted human bodies in yoga images. Therefore, the inpaint function is usually needed to further correct image details.
In addition, a month ago, Adobe released the news of Firefly, which attracted the attention of the entire technology field. The birth of Firefly marks the official entry of products in Adobe's “family bucket” into the AIGC field, and this is just the beginning.
Compared with AI drawing technologies such as Stable diffusion and Midjourney, Adobe Firefly’s products are of greater significance. It can output ultra-high-resolution content, and then layer and change details.
Initially, Firefly’s goal was only to generate commercially safe static images, but it has now announced the introduction of Adobe’s video, audio, animation and dynamic graphic design applications. In the future, everyone can easily edit videos with just a few words to automatically color grade, add background music and sound effects by using it.
It is worth noting that the above effects are only Adobe’s introduction of these functions. The actual use effect needs feedback from users after real experience.
The development of AI image processing technology is changing with each passing day, but there is still a long way to go to become a complete production tool. There are some AI companies that focus on technical research but are rarely noticed by people, such as JUNLALA, a recently emerging artificial intelligence company, which has been committed to breakthroughs in AI image processing technology since its establishment.
As one of the important players in the AIGC era, JUNLALA is expected to bring amazing products to people in the near future with its long-term technical accumulation. The JUNLALA team aspires to become a global leader in the field of artificial intelligence technology innovation, and always regards optimizing user experience as its core mission. Technology has changed the world and people's work efficiency. Life and entertainment experience have been improved unprecedentedly through AI teams such as JUNLALA. A more colorful picture of life will slowly unfold in the human world.