On the morning of May 14th (Vietnam time), the American company OpenAI launched GPT-4o - a new version of its ChatGPT artificial intelligence (AI) software, capable of interacting via text and images, as well as voice chat.
A friend who understands the user's emotions.
Since its launch in late 2022, and after several upgrades, ChatGPT is now popularizing versions ChatGPT3.5 and ChatGPT 4 (Plus) to support users in handling daily tasks by inputting questions to receive results.
GPT-3.5 is currently the most widely used version due to its free availability, offering impressive processing capabilities such as writing fluent speeches quickly, solving math problems with high accuracy, and analyzing data comprehensively. This is a task that would normally take users hours, even days, to complete. GPT-4, the paid version at $20/month (approximately 500,000 VND), surpasses GPT-3.5 and is considered "super intelligent" by users thanks to features such as: reading and understanding text, analyzing and processing attached Excel and PDF files; analyzing and editing images using AI commands; drawing images directly with AI; drawing charts and other complex tasks in a single operation…
These two versions have satisfied and attracted hundreds of millions of users. To provide even more convenience and user experience, OpenAI surprised everyone by launching a new model called GPT-4o ("o" stands for omnipotent). Users can use the latest version for free, and paid users can use it five times more frequently. The highlight of GPT-4o is its more natural and user-friendly interface compared to the GPT-4 model. In the new version, the application can remember the context of conversations, learn from previous interactions with users, and provide real-time multilingual translation. GPT-4o enhances interaction through text and images, as well as voice chat.
Furthermore, the new version of ChatGPT can communicate using natural-sounding human or robotic voices, even singing a part of a response. It also has the ability to recognize user emotions: listening to breathing patterns and offering advice to help users calmly resolve issues. Notably, this tool can analyze images and charts and discuss based on information provided by the user. Even more impressively, with integrated translation and automatic reply features, this version supports real-time multilingual conversations (over 50 languages). With this feature, the tool eliminates language barriers, becoming a friend to users, allowing for continuous, uninterrupted conversation. To increase user accessibility, OpenAI has also released a desktop version for macOS, integrating the features of the GPT-40 model, offering users another platform to choose from. Accordingly, users can now ask quick questions using keyboard shortcuts, chat via voice, and discuss screenshots. The Windows version is expected to launch this year.
According to several sources, the enhanced ChatGPT 5 version will be released soon, featuring significant upgrades in both performance and functionality. This new version will focus on improving natural understanding and interaction, optimizing security, and providing a smoother user experience.
Users can use the GPT-40 version for free and pay to use it five times more frequently. (Photo: Le Tinh)
Security should not be neglected.
Alongside the unlimited development of AI software like ChatGPT, users need to be aware of security issues related to personal data.
Specifically, set strong passwords with a minimum length of 12 characters and a variety of different character types, along with 2-3 layers of protection for your account and recovery email to minimize the risk of account loss. In addition, before installing web browser extensions on your computer, thoroughly research whether those applications are developed by reputable programmers. For e-wallets or bank accounts used to pay for paid versions like ChatGPT, users should use the "block card" feature online after payment or only maintain a sufficient balance to avoid financial risks. Furthermore, when wanting to use new AI software, users need to carefully research the publishers and distributors of AI software activation accounts to avoid losing money when purchasing applications online.
Furthermore, be cautious of links and attachments about new AI software sent by strangers, as these links may contain malware designed to infiltrate your device and steal information. Look for AI software on official app stores such as the App Store, Google Play, Google Play Store, Chrome Web Store, or directly from the official website or reputable distributors. Finally, ensure a secure internet connection and that your browser and operating system are up-to-date.
Gemini 1.5 Pro to be released soon.
Recently, Meta Platforms – Facebook's parent company – announced versions of Llama 3, a model that will be integrated into the Meta AI virtual assistant. This new AI provides performance comparisons on topics such as reasoning, programming, and creative writing, allowing users to perform tasks within Meta's data repository, such as finding recipes, reading news, or performing simple searches, instead of being forced to exit Instagram, Facebook, or WhatsApp. In March, Google announced the upcoming release of Gemini 1.5 Pro, a superior version of Gemini 1.0, capable of capturing 1 hour of video , 11 hours of audio, and over 30,000 lines of code via user prompts. Users can then input entire books, legal case histories, or anything else they want into Gemini and receive answers to questions based on accurate data.
Province
Source: https://nld.com.vn/dung-chatgpt-toi-uu-an-toan-19624051420381976.htm
Comment (0)