laitimes

Ren Zixing: It is expected to take advantage of the ChatGPT plug-in era and develop plug-in + security protection

author:Palm stock news

Plugins released, ushering in the AIOS super era. OpenAI has announced the addition of plugin support for ChatGPT, which allows ChatGPT to access other third-party websites and related applications and present a simple one-stop operation. At present, the official website shows that the plug-in mall has been directly connected to more than 10 applications, and the plug-in has three basic functions (retrieving real-time information, retrieving knowledge base information, and operating applications instead of users), and the development and call are extremely convenient. We believe that the release of Plugins means the arrival of the AIOS era, and there will be three far-reaching impacts in the future: 1) All kinds of applications combine a hundred flowers, the application super era, and may even "follow the word", directly form plug-ins through language, and quickly change the physical world. 2) In the Plugin ecosystem, large model capabilities will be very important, AIOS is both a system and an entrance, with the potential of large model capabilities, and technology manufacturers with excellent ecological capabilities will be more important. 3) It is expected that plugins will make it easier for users to link with the underlying IT, and the vendor barriers with the underlying IT system are highlighted, and can become the provider of key plug-ins. For example, manufacturers with the underlying information system of the hotel can become direct suppliers of plug-ins. If permissions permit, user bookings can be made directly through ChatGPT combined with plugins.

The official website shows that the plug-in mall has been directly connected to more than 10 applications. The first plugins for ChatGPT were provided by companies such as Expedia, FiscalNote, Instacart, Kayak, Klarna, Milo, OpenTable, Shopify, Slack, Speak, Wolfram and Zapier, and the specific functions of these plugins include recommending restaurants, making online shopping, travel, office, searching, language teaching, etc., covering all aspects of clothing, food, housing and transportation. The launch of the ChatGPT Plugins Store means that other services have become tools. For example, the OpenTable plugin will allow chatbots to search for available reservations in restaurants, while the Instacart plugin will allow ChatGPT to place orders from local stores.

$Ren Zixing (SZ300311)$

Plugin generation development will be one of the best businesses in the era of ChatGPT plugins:

Ren Zixing: It is expected to take advantage of the ChatGPT plug-in era and develop plug-in + security protection

A plug-in management method for web page collection tasks and its device, the implementation of the present invention, by combining the advantages of Java scripting language, as well as the existing version control server, improve the development efficiency of plug-ins, realize the automatic management of plug-ins, thereby greatly reducing the cost of plug-in development and plug-in maintenance. At the same time, during plug-in development, it is possible to support dynamic repeated loading of plug-ins without restarting the program, and the management operation is more secure and convenient when managing plug-ins.

A plug-in development method and system, the present invention discloses a plug-in development method and system, the method comprising: S1, input URL address and convert to DOM tree; S2. Get Java script extraction rules and generate plugins; S3, test plug-in; S4. Test whether the plug-in passes, if it passes, go to S5, if not, go to S2; S5. Upload and store the plug-in. The beneficial effect of implementing the present invention is, quickly, efficiently, and safely develop the plug-in.

Improve the development efficiency of plug-ins, realize automatic management of plug-ins, thereby greatly reducing the cost of plug-in development and plug-in maintenance, and solve plug-in development and management.

Ren Zixing: It is expected to take advantage of the ChatGPT plug-in era and develop plug-in + security protection

We believe that the release of Plugins means the arrival of the AIOS era, which will have three profound implications in the future:

The combination of various applications and ChatGPT is more flexible, which can quickly enable CHATGPT to empower all things intelligently, regardless of the type of application. ChatGPT's Codex capabilities may even make it possible for users to create plugins by way of description to directly invoke physical or software applications, quickly changing the physical world. For example, smart headsets directly call WeChat, documents, Dianping, Baidu maps, car music software and other applications, smart speakers and various smart home quick links, call air conditioners, TVs, sweepers, computers, dishwashers, smart door locks, smart lamps, etc.

Under the plugins ecosystem, the big model has a strategic position similar to the operating system, and it is also the core entrance for future traffic, users, data, etc. Technology manufacturers with large model capabilities and excellent ecological capabilities will be more important.

IT IS EXPECTED THAT PLUGINS WILL MAKE THE LINK BETWEEN USERS AND THE UNDERLYING IT MORE CONVENIENT, AND THE VENDOR BARRIERS WITH THE UNDERLYING IT SYSTEM WILL BE HIGHLIGHTED, AND IT CAN BECOME THE PROVIDER OF KEY PLUG-INS. For example, in the hotel scenario, a manufacturer with a hotel's underlying information system can become a direct supplier of plug-ins. Where permissions allow, user bookings can be made directly through ChatGPT combined with PLUGINS.

$Renzixing (SZ300311)$ The zero-trust AI solution facilitates and fast user operation, intelligently prevents data leakage, and comprehensively protects the data security of customers' important and sensitive business systems. Image persistent storage provides data support for accident traceability and forensics, realizes the closed-loop of scene application, and has great product value.

【Artificial intelligence first team Minsheng computer】ChatGPT failure leads to data leakage, pay attention to the algorithm and data security industry chain represented by [360].

Event grooming:

1) On March 20, due to Redis' open-source library bugs, ChatGPT malfunctioned and data leaked, causing some users to see other users' personal information and chat query content;

2) On March 22, OpenAI CEO Sam Altman apologized on Twitter for the leak. Altman tweeted: "We have had a major issue with ChatGPT due to a vulnerability in the open-source library, and now that a patch has been released, we have just completed validation work. Only a small percentage of users will be able to see the title of other users' conversation history. We apologize for this. ”

3) On March 25, OpenAI released a post-mortem analysis report explaining that a vulnerability in the Redis customer software open-source library caused the ChatGPT service to leak other users' chat queries and the personal information of 1.2% of ChatGPT Plus users. In a postmortem analysis released today, OpenAI said: "The vulnerability was discovered in Redis-py, the open-source repository for Redis customer software. As soon as we discovered the vulnerability, we contacted the Redis maintainers and a patch was released to address the issue. "The leaked information included the subscriber's name, email address, payment address, last four digits of the credit card number, and expiration date.

Based on identity, the Zero Trust solution adheres to the principle of "network unprivileged, trust minimization, and dynamic permissions", adopts software-defined boundary technology, strengthens identity governance and access control, makes full use of situation awareness, traffic analysis, asset monitoring, behavior portrait, and combines application stealth and terminal access and control to continuously and dynamically build security protection barriers for enterprise core assets from the "cloud", "management" and "end".

The new function of the product is the successful application of AI technology, which lays the technical cornerstone for the expansion of other intelligent functions of the product.

This new feature facilitates quick and easy user operation, intelligently prevents data leakage, and comprehensively protects the data security of customers' important and sensitive business systems. Image persistent storage provides data support for accident traceability and forensics, realizes the closed-loop of scene application, and has great product value.

Based on the forward-looking layout in the field of zero trust security and the acceleration of social digital transformation, Renzi Bank has now formed a zero trust security solution for customers in various industries with multiple application scenarios, adhering to the principle of "never trust, continuous verification", through comprehensive perception, continuous evaluation, dynamic authorization and other technologies to help customers achieve secure access in multiple application scenarios such as remote office and business access.

The distress of ChatGPT or generative AI

The impact of AI is both good and bad. For example, some cybercriminals are already using ChatGPT to write malware, and ChatGPT or generative AI may pose the following potential threats:

Phishing is a means by which cybercriminals steal confidential information such as online banking login accounts and passwords, credit card information, business login credentials, passwords/passphrases, etc. by sending fraudulent information. Phishing is usually email-based, but the scam has now expanded beyond suspicious emails to phone calls (so-called "phone scams"), social media, text messages (also known as "SMS scams") and apps.

AI and machine learning solutions may be good at identifying trends in social engineering and malicious content, but when it comes to sender identities, they don't provide much useful information because email attackers change identities quickly.

In the face of the AI wave, we take precautions

Currently, more than 800 AI policy initiatives in 69 countries are on the way, but the application of generative AI models has not yet been involved. Existing and proposed AI regulations cover several specific use cases (e.g., data privacy, discrimination, surveillance, etc.) as well as specific decisions (e.g., hiring, lending, website referrals, public contracts, etc.), and most of these regulations are developed in response to AI's potentially harmful effects on people and society.

The top threat to cybersecurity in 2023 is likely to be criminal services; Criminals can offer their services on these platforms to those who would otherwise not be able to carry out cyberattacks. Since ChatGPT is able to speed up the malware creation process without charging anything, this could make criminal services more profitable for cybercriminals. Criminal software businesses will continue to threaten businesses of any size. They can help entry-level cybercriminals commit crimes, which earn more and are less risky than committing crimes directly against victims.

Between positive and negative outcomes, there has been a race between the development of AI technology. In the face of the booming wave of AI, we should let it play its due value and reduce its negative effects. If built and used responsibly, AI will create prosperity and enrich people's lives. Safety should be the primary prerequisite for the development of AI.

Ren Zixing: It is expected to take advantage of the ChatGPT plug-in era and develop plug-in + security protection