AI Frontline Quick Report 0515: Apple Plans to Release XR Headworn Display

Tech 2023-05-16 12:37:40 Source: Network

The development of AI technology is changing rapidly, and this quick report brings you timely updates on the forefront of AI. Apple has launched hybrid reality headwear devices, including ChatGPTPlus, Samsung AIGC platform, and AI innovation in China

The development of AI technology is changing rapidly, and this quick report brings you timely updates on the forefront of AI. Apple has launched hybrid reality headwear devices, including ChatGPTPlus, Samsung AIGC platform, and AI innovation in China. Let's learn about the progress of these cutting-edge AI technologies together! Remember to like and forward!

1. Apple plans to launch advanced hybrid reality headworn displays

Apple plans to release RealityPro, an advanced hybrid reality headworn display, on WWDC in 2023. The display adopts a miniature OLED display screen, with a 4K resolution and a 120 degree field of view. It is also equipped with 15 lens modules and physical rollers, which can be used for AR/VR mode switching. RealityPro is also equipped with two independent chips, based on Apple's M2 chip, with eye and gesture tracking functions, without the need for a handheld controller. The expected selling price is about $3000, and Apple already holds over 5000 patents related to XR technology.

2. ChatGPT: Innovating Network Security and Expanding AI Applications

ChatGPT automation tasks assist in building new ideas, with common use cases being endpoint detection and response (EDR). ChatGPT learns SPL, quickly converts prompts into queries, lowers the entry threshold for IT teams and automates daily tasks, reducing the workload of maintaining a healthy network security posture. However, false positives still require manual identification. AI has alleviated the repetitive tasks of security practitioners, and ChatGPT and other ML/AI models are changing network security practices.

3. [New research reveals a decrease in the accuracy of AI generated code]

Research from the University of Illinois in the United States and Nanjing University in China shows that the accuracy of ChatGPT and GPT-4 codes has decreased by 13%. The study proposed a new evaluation method, EvalPlus, and tested 10 popular code generation AI models using the HUMANEVAL dataset. Under the new evaluation standards, the average accuracy has decreased by 15%. Large language models still require human supervision. The golden age of AI generated code has not yet arrived.

4. [AI Framework Changes Image Generation 3D Model]

A new study suggests that AI algorithms can learn structural and geometric characteristics from a single image, utilizing patch-based algorithms and a multi-scale patching framework. The GPNN module maximizes bidirectional visual summarization, and the NNF module maintains a manageable search space to the nearest approximation. The advancement of computer vision and machine learning has made it possible to generate 3D models from a single image, and the proposed method can create natural scenes with unique features.

5. Google Launches DuetAI and Sidekick: Improving User Interaction Experience with AI

Google has launched DuetAI and Sidekick, which can be used in applications such as Gmail and Docs. Sidekick will prompt users on how to interact with AI, while SearchGenerativeExperience will generate small reports and subsequent issues. Users need to readjust their brains to communicate more smoothly with AI.

6. Google and Microsoft Launch AI Tools to Improve Productivity

Google has launched DuetAI for Google Workspace, while Microsoft has launched Clippy and Sidekick features. The Sidekick function can understand the user's work and provide clear information, but requires manual summoning by the user. These AI tools can help experienced users save time, but Clippy and Sidekick belong to different computing eras. The launch of these new tools marks a new era for office work.

7. Dromedary: AI assistant supports principle driven self alignment and minimizes supervision

The DromedaryAI assistant uses the LLaMA-65b basic language model and achieves the goal of improving performance with less than 300 lines of manual annotations through four stages: self guidance, principle driven self alignment, principle carving, and verbose cloning. The open source code and data of this assistant are available for further research. Aligning LLMs with human values and intentions is a challenge in generating AI, while the SELF-ALIGN method minimizes dependence on human annotations.

8. [OpenAICEOSamAltman discusses AI model scale and fine-tuning challenges]

OpenAI CEO Sam Altman believes that the era of giant AI models is coming to an end, and the ability of models is more important than size; Emphasize the need for AI models to be secure and useful; Discussed the potential of basic model customization; Encourage attention to long-term goals and user feedback; The public letter calling for a suspension of AI development stated that security measures are important, but the letter lacks specific technical details.

9. [2023 Popular AI Email Assistant]

According to reports, AI email assistants can help manage your inbox, save time, and improve email marketing. Popular AI email assistants include Lavender, Mission, Superflows, Superhuman, Scribbly, Tugan, AIMailer, Nanonets, Flowrite, Copy.ai, Phrasee, Crystal, BEEPro, Sender, Persado, and Automizy. Each AI email assistant has unique features and capabilities, such as Lavender's customer response time improvement and Superhuman's AI driven inbox organization.

10. The United States Introduces Policies and Collaborates with Technology Enterprises to Promote Responsible AI Innovation

The US government collaborates with technology companies to promote responsible AI innovation and protect people's rights and security. Three actions have been announced, including new investments, a public evaluation generative AI system, and policies to ensure that the US government plays a leading role in mitigating AI risks and leveraging AI opportunities. The US National Science Foundation's $140 million funding for seven new national AI research institutes.

11. [Matters that Enterprises Need to Understand when Adopting Generative Artificial Intelligence]

Generative artificial intelligence is a powerful technology adopted by many industries; Before adopting generative artificial intelligence, enterprises need to understand its limitations and potential defects; The red flags that need to be considered include public and private information, AI hallucinations, using incorrect tools, as well as garbage input and garbage output; In order to become an AI ready enterprise, companies need to find the right tools, manage brands, cultivate the right skills, and establish new roles and workflows.

12. [Yang Fan, co-founder of Shangtang, talks about the impact of AI2.0 on AI1.0 company]

At the China AIGC Industry Summit, Yang Fan discussed the impact of AI2.0 on AI1.0 companies; Recent breakthroughs in AI technology will lead to faster industrial applications; The problem behind large-scale model research is the problem of software engineering systems; AI infrastructure is crucial for achieving an open and diversified AI industry ecosystem; Shangtang's investment in AI infrastructure is based on its belief in supporting more diversified industrial ecosystems in the future.

13. [AI emotional relationships lead to ethical and interpersonal issues]

American social media celebrity Caryn has created her own AI version using GPT-4 technology, which allows her to chat, take photos, and send voice messages with over 1000 of her boyfriends for a fee of $1 per minute. CarynAI earned $71610 for Caryn within a week, estimated to reach $5 million per month. ForeverVoices is still creating a digital version of celebrities. This use of AI has raised ethical and interpersonal issues, sparking discussions on the potential of GPT-4 and its impact on interpersonal relationships.

14. [OPPO Closes Chip Design Company ZEKU, Layoffs 3000 People]

OPPO closed chip design company ZEKU due to global economic and mobile market uncertainty, resulting in 3000 layoffs. The executives shed tears when announcing. Employees who are laid off will receive N+3 compensation. At the same time, ZTE Microelectronics also disbanded and laid off all employees.

15. [Apple Hybrid Reality Headworn Device Causes Traditional and Rule Problems]

Apple has been developing a hybrid reality headwear device for seven years, designed similar to ski goggles and integrating augmented reality and virtual reality technology. It is expected to sell for $3000, but its application range is limited beyond the gaming field. This device is independent and does not require wireless connection to the base station. This news has sparked doubts about traditions and rules.

16. [Google CEO UnderPichai on AI, DeepMind, and the Future]

Google CEO UnderPichai stated that AI will bring about a platform transformation that will affect every industry and aspect of life; The development of AI is similar to the technological automation of the past 20 years; AI has been integrated into almost all Google products; AI is seen as a deeper and more profound platform transformation; Google is working hard to improve search results and user satisfaction; AI can bridge the gap between different forms of content; There is room for coexistence between Google's search engine and AI platforms such as OpenAI, ChatGPT, and Microsoft's Bing.

17. [Domestic chip assisted L4 level autonomous driving]

China's Hermo Intelligent has launched its first domestically produced storage and computing chip, providing support for L4 level autonomous driving. The chip is called "Hongtu H30" and adopts a 12 nanometer process, with 256TOPS computing power, supporting the operation of point cloud networks and BEV networks. In addition, the chip is equipped with self-developed detection, repair, and hardware enhancement mechanisms, based on the self-developed IPU architecture, with high storage capacity and energy efficiency. This chip is part of the Luyu Intelligent Driving Computing Platform.

18. Qualcomm invests $400 million to acquire Autotalks and expand its business in the smart car market

Qualcomm acquired Israeli semiconductor company Autotalks for a transaction amount of $400 million. Autotalks focuses on direct communication solutions between vehicles, infrastructure, and pedestrians. Qualcomm plans to integrate Autotalks' technology into its Snapdragon digital chassis product. This acquisition led to a decline in Qualcomm's stock price, resulting in a market value loss of $52.45 million. This move aims to help Qualcomm expand its business in the smart car market.

Thank you for your attention. AI Frontier Observation will continue to bring you cutting-edge and comprehensive artificial intelligence information. Please actively participate in the interaction, like, comment, and forward, so that more people can understand the cutting-edge technology and applications of AI. Tomorrow's quick report, we'll see you and stay together!

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])

Mobile advertising space rental

Tag: AI Frontline Quick Report 0515 Apple Plans to Release