According to the latest news, DeepMind co-founder Mustafa Suleyman recently dropped a bombshell in an interview: OpenAI is secretly training the GPT-5, which will be 100 times larger than the current GPT-4 in the future.

The era of the GPT-5 model is coming

At the beginning of the year, OpenAI officially released GPT-4. Compared with the GPT-3.5 model originally used by ChatGPT, GPT-4 achieved several leaps: strong map recognition ability; text input limit increased to 25,000 words; answer accuracy significantly improved; can generate lyrics and creative text to achieve style change.

GPT-3.5 has a parameter scale of 175 billion, and OpenAI has not announced the specific specifications of GPT-4, but later analysis concluded that it has 16 expert models, nearly 1.8 trillion parameters, and 13 trillion training data. If GPT-5 is upgraded at this ratio, it will exceed the 10 trillion parameter and could become the largest AI model, far more than any other competitor.

If the previous GPT-4 can only be used as a novel and interesting toy, it can write an email, chat, and complete some boring tasks. Then, the enterprise-oriented GPT-5 will change the user working scene and change the existing working mode, which means that the strongest AI GPT-5 will finally debut, and the era of man-machine collaboration is coming.

In the first half of 2023, the big model-related track herd. To some extent, who can be the first to seize a place on this track, will master the password of the next era.

Known as the “slow half beat” Apple (AAPL), also has a new trend. Mr Cook revealed that Apple’s R&D spending on generative AI will increase further in the coming months. Sources say Apple has built its big language model framework, Ajax, and applied it to maps, Siri, and other features.

For Google (GOOG), the upcoming new Pixel model will incorporate the latest PaLM 2 big model gecko version. The Information Exclusive reports that Google’s multimodal AI model, Gemini, is coming out, directly against OpenAI’s GPT-4. There are rumors that the model will have at least 1 trillion parameters and be trained using tens of thousands of Google TPU AI chips.

WiMi Hologram Cloud seizes the opportunity of GPT

The big model pattern may still rise at the moment. In the “changing” environment, only grasping the forward trend, and adhering to the core competitiveness is the “constant” law to deal with the change. Public information, WiMi Hologram Cloud (NASDAQ: WIMI) as a “leader in the field of AI vision”, over the years always adhered to the long-term strategy, continued to the important frontier research in the field of artificial intelligence, and took the lead in fully embrace AIGC, seize the opportunity of AIGC technology, continue to upgrade existing products, expand new business, create diversified AI + integrated services, provide users with more rich and professional consumption experience.

In the context of the current vigorous development of generative artificial intelligence, WiMi Hologram Cloud is constantly improving the quality of industrial development and the advantages of the whole industrial chain and is committed to promoting the productivity transition of the digital creative industry. In this regard, WiMi Hologram Cloud focuses on cultivating and expanding emerging industries, including advertising, retail, education, automobile, and other traditional industries equipped with AI, yuan universe, and other new digital technologies to carry out cross-border integration, improve the industrial scientific and technological innovation ability, create new competitive advantages, and bring new great value to them.

In addition, given the challenges brought by large models, the continuous linkage of large models and algorithmic models also requires continuous investment and innovation in the underlying computing power. WiMi Hologram Cloud Closely coordinated ecological partners, industry-leading enterprises, and scientific research talent depth cooperation, give full play to the administration, production, study, and research, with all strength, build a big model for different application scenarios solutions, from the underlying calculate force, framework, algorithm, application full stack AI algorithm ability, the future will meet the demand of multiple scenarios, accelerate the big model to explore innovation and industrialization to the ground.

To sum up

Since ChatGPT has become popular around the world, the AI model has attracted the attention of several leaders in the technology circle and brought a comprehensive outbreak of generative AI. According to media reports, the Ministry of Industry and Information Technology and the Ministry of Finance jointly issued the Action Plan for Stabilizing Growth of the Electronic Information Manufacturing Industry 2023-2024, proposing to encourage the construction of data infrastructure and AI infrastructure to meet the needs of the application of artificial intelligence and big model.

With favorable policies, the expansion of application scenarios, and the continuous progress of technology, the big model will achieve commercialization in more fields and promote industrial upgrading and reform. The market demand for generative AI applications is strong, which injects a “shot in the arm” for the landing of AI applications, and the AI + track welcomes good news. Therefore, in this trend, the big model track players also need to accelerate the expansion of their own business scope and service areas, to meet the needs of different industries and customers.