Meta plans to launch its first AR glasses in 2024
According to the latest report from The Information, Meta is expected to launch its first true AR glasses, codenamed "Orion," in 2024. But due to cost issues, the company plans to produce only about 1,000 AR glasses for internal development and presentation to the public.
Leia introduces the one-click to 3D animation tool
Leia is a leading US-based supplier focused on glasses-free 3D display hardware and software solutions. Leia has launched LeiaPix Converter, an AI-powered image processing tool that converts static 2D images into dynamic 3D images.
Currently, LeiaPix Converter has more than 1 million users. At the end of July, Leia announced that it would significantly expand the capabilities of LeiaPix Converter and provide cloud API support for its own 3D tablets. LeiaPix Converter is currently available online for free.
Deloitte Japan will use AI to review corporate financial information
Deloitte Tohmatsu will officially begin using artificial intelligence (AI) to review companies' financial information as early as September. AI learns the financial reports of companies that have made mistakes and misconduct in the past to identify the risk of non-compliance of audited companies.
Microsoft: GPT-4 is good enough for medical tasks
A Microsoft study concluded that GPT-4 is sufficient for medical tasks, speeding up medical processes and improving efficiency. According to the Microsoft research team, large language models (LLMs) such as GPT-4 have great potential in the medical field. These models can help speed up medical processes, for example by processing large-scale unstructured patient data to improve the efficiency of cancer drug development.
Microsoft has also introduced language models such as BioGPT, which is specifically for medical tasks, but has now made it clear that it will rely heavily on GPT-4 in the future. In addition, Microsoft is developing the "LLaVA-Med" model to more closely integrate medical data with research, accelerating medical care and research.
OpenAI could go bankrupt in 2024
According to a recent report by Analytics India Magazine, ChatGPT services cost around $700,000 per day. Although OpneAI has been trying to make a profit with GPT-3.5 and GPT-4, it has not yet generated enough revenue and has not broken even. While the prospects for commercial revenue are not satisfactory, ChatGPT's user base is also gradually declining, and in July, the product's user base plummeted by 12%, from 1.7 billion to 1.5 billion.
In addition, generative artificial intelligence launched by other vendors, as well as some open source models, have also had an impact on the OpenAI market. Under the combined impact of these factors, if there is no more capital injection, OpenAI is likely to be forced to file for bankruptcy before the end of 2024.
Domestic large model evaluation: Wen Xin ranked second in a word
The China Enterprise Development Research Center of Xinhua News Agency Research Institute released the "Artificial Intelligence Large Model Experience Report 2.0" yesterday. The "Report" pointed out that the current domestic large model products have made significant progress, but compared with humans with higher education, there is still a certain degree of gap in IQ, emotional quotient and other aspects.
The "Report" selects 8 large model products of 360 Wisdom Brain, Baidu Wenxin Yiyan, Lanzhou Mchat, SenseTime, Xunfei Xinghuo, Ali Tongyi Qianwen, Kunlun Tiangong, and ChatGLM for evaluation, and calculates the total score according to four dimensions: basic ability, IQ ability, emotional intelligence ability, and tool extraction. From high to low, the rankings are iFLYTEK Xinghuo, Baidu Wenxin Yiyan, SenseTime, ChatGLM, 360 Wisdom Brain, Kunlun Wanwei Tiangong, Ali Tongyi Qianwen, and Lanzhou Mchat.
According to reports, iFLYTEK Xinghuo "has obvious advantages in improving work efficiency", Baidu Wenxin said that "the foundation is deep and the basic ability is still at the leading level", SenseTime Discussion is "excellent in emotional intelligence", and the overall performance of ChatGLM of Zhipu AI is excellent. At the same time, the focus of AI technology development has shifted from the blind pursuit of large models to the development of practicality and sustainability, and it takes more time to verify the large-scale implementation of AI.
For more metaverse information, please pay attention to the public account "Metaverse Chronicles"
Reply to "Reports" to get the latest and most authoritative industry reports