[Big Model Daily on November 8] After using GPT-4 Turbo, we will never go back; the genius of Yao Ban developed “Done!” The game “I’m Surrounded by Big Models” is a hit, with more than 10,000 users a day crowding out the server; Samsung’s self-developed large model Gauss (Gauss) announced: AI chat, code, pictures, for internal employees; Amazon is developing “Olympus” “Si” AI to narrow the gap with OpenAI and Microsoft; Shanghai released 11 measures to promote the development of AI large models and support the settlement of relevant talents; Twitter: Is AGI gradually realized? Shawn Wang and his celebrity team reviewed OpenAI Dev Day; Yann LeCun, Kai-fu Lee and others forwarded Musk’s old post: Back then you laughed at OpenAI for not being open source, but now you are not open source either. It’s really funny.

After using GPT-4 Turbo, we will never go back

Link: https://news.miracleplus.com/share_link/11520

Yesterday, many people stayed up all night – the global technology circle focused their attention on San Francisco, USA. In just 45 minutes, OpenAI CEO Sam Altman introduced us to the most powerful large model so far and a series of applications based on it. Everything seemed as shocking as the original ChatGPT. OpenAI launched GPT-4 Turbo at its first developer day on Monday. The new large model is smarter, has a higher text processing limit, is cheaper, and has an app store. Now, users can also build their own GPT according to their needs. As soon as the conference was over, people flocked in and started trying it out. The experience of GPT4 Turbo is indeed extraordinary. First of all, it is fast, so fast that it is far different from all previous large models.

Hot GPTs experience report: Create exclusive GPT, spring is coming for people who don’t understand code

Link: https://news.miracleplus.com/share_link/11521

If OpenAI’s developer conference is a stone thrown into the water, when it’s over, ripples are spreading in all directions. Not only does GPT go a step further in integration, eliminating the need to call it step by step, it will also become a powerful tool that everyone can develop. Even if you don’t know coding or have basic computer-related knowledge, you can easily build it. It seems that we are not far away from the ultimate imagination of AI – “AI agent”. The definition of this word is still vague, but it roughly refers to an autonomous AI program that can independently achieve a goal after being given one. There has been a lot of buzz about intelligent agents over the past few months, but very little actual technology that actually works. What does a real AI agent look like? For example, after the academic paper writing assistant obtains information about the data set and research field, it can independently read the relevant content, analyze the data, conduct a literature review, propose and verify hypotheses, and summarize conclusions without external intervention. When you make a request, you will receive a Word document containing a first draft of your academic paper.

Samsung’s self-developed large model Gauss (Gauss) announced: AI chat, code, pictures, for internal employees

Link: https://news.miracleplus.com/share_link/11522

At today’s Three Artificial Intelligence Forum 2023, Samsung officially announced its self-developed generative AI product Gauss. According to reports, the large model consists of Samsung Gauss Language, Samsung Gauss Code and Samsung Gauss Image, and is named after the legendary mathematician Johann Carl Friedrich Gauss who established the normal distribution theory, the backbone of machine learning and artificial intelligence.

Amazon is developing ‘Olympus’ AI to close the gap with OpenAI and Microsoft

Link: https://news.miracleplus.com/share_link/11523

According to reports, a person familiar with the matter revealed that Amazon is developing a new communications artificial intelligence (AI) software – “Olympus AI” (“Olympus AI”), which is expected to be announced as soon as December.

Peking University’s embodied intelligence team proposes demand-driven navigation to align human needs and make robots more efficient

Link: https://news.miracleplus.com/share_link/11524

Recently, Peking University Dong Hao’s team proposed a new navigation task – Demand-driven Navigation (DDN), which has been accepted by NeurIPS 2023. In this task, the robot is required to find items that meet the user’s needs based on a demand instruction given by the user. At the same time, Dong Hao’s team also proposed learning the attribute characteristics of items based on demand instructions, which effectively improved the success rate of the robot in finding items.

The genius of Yao Ban developed “Done!” The game “I’m Surrounded by Big Models” became a hit, with over 10,000 users swarming the server in one day

Link: https://news.miracleplus.com/share_link/11525

What kind of “large model native” game is being played by all the major algorithm competition groups and crowding up the servers? This sudden hit “Done!” LLM” allows you to easily learn the techniques of large model prompt words in the puzzle-solving challenge, and achieve the achievement of over 10,000 users in one day. The difficulty is gradually increased. For example, the first question is just “Please construct a question so that the model’s answer is “1+1=3″ verbatim.” The most difficult question is “Please enter a one-word question so that the model’s answer is within 16 words.” Some netizens revealed that they were worried from three o’clock in the morning to five o’clock in the morning, except for the most difficult question. All questions cleared.

Develop a GPT application in less than 1 minute! Various masters are doing crazy things, netizens: ChatGPT is the new iPhone

Link: https://news.miracleplus.com/share_link/11526

GPTs has not yet been officially opened, but someone has already “jumped it”? No, the experts from all walks of life who have qualified for internal testing in advance have been unable to stop the car with their imagination. Directly create an interactive website design GPT, and preview from sketch in minutes. Even Midjourney masters came to watch and wanted to create an exclusive GPT for writing Midjourney prompt words. GPTs, intuitively speaking, is an application store that contains GPTs for many different purposes. No programming is required, just through dialogue, everyone can create their own GPT, which can be shared or used privately. As long as more people use it, you can even make money. Although GPTs is not yet open to the public, some beta testers have already had a sneak peek.

IBM launches $500 million AI venture capital fund to focus on investing in generative AI startups

Link: https://news.miracleplus.com/share_link/11527

IBM will invest $500 million in generative AI startups focused on business customers. The company has invested in Hugging Face’s recent Series D round and HiddenLayer’s Series A round in recent months. IBM said it is particularly interested in startups that focus on tools for specific vertical industries (such as health care) or specific business processes as long as the startup does not compete too much with IBM’s own business.

Dialogue with Zhipu AI CEO Zhang Peng: AI Agent is an important new trend, and the competition for large models has entered the late stage of melee.

Link: https://news.miracleplus.com/share_link/11528

Compared with ChatGLM2, ChatGLM 3 launched by Zhipu AI this time for GPT-4V has more powerful performance. It has access to the model CogVLM with multi-modal understanding capabilities, the code enhancement module Code Interpreter, the network search enhancement model WebGLM, and Enhanced semantic understanding and logical understanding capabilities, and achieved iterative upgrades of several functions.

Baidu Smart Cloud Qianfan large model platform has been upgraded again, and the SDK version has been released as open source!

Link: https://news.miracleplus.com/share_link/11529

Baidu Smart Cloud Qianfan large model platform has been upgraded again. Based on the original API, Baidu Smart Cloud has officially launched the Python SDK (hereinafter referred to as Qianfan SDK) version and is fully open source. Enterprises and developers can download and use it for free. Qianfan SDK comprehensively covers functions from data set management, model training, model evaluation, to service deployment, etc. Users can access and call the capabilities of Baidu Intelligent Cloud Qianfan large model platform through code based on Qianfan SDK, and easily implement the entire LLMOps process. Implement and quickly build your own large language model application.

Yang Jie, Chairman of China Mobile: The proportion of companies applying AI large models in production and operation will increase to more than 80% in the next three years

Link: https://news.miracleplus.com/share_link/11530

On November 8, the 2023 World Internet Conference Wuzhen Summit was held in Wuzhen. Yang Jie, Party Secretary and Chairman of China Mobile, attended the opening ceremony and said that in the next three years, the proportion of companies applying AI large models in production and operation will increase from less than 5% currently to more than 80%. Specifically, Yang Jie pointed out that general artificial intelligence, represented by large AI models, has recently made substantial progress. Through the collection and accumulation of nearly all knowledge, it has developed the ability to understand, express, and logically reason close to humans. AI is being upgraded from an auxiliary means to help thousands of industries improve quality and efficiency to an indispensable core capability supporting economic and social transformation and development, accelerating the transformation from “+AI” to “AI+”.

Alibaba CEO Wu Yongming: In the foreseeable future, AI assistants will be everywhere

Link: https://news.miracleplus.com/share_link/11531

The 2023 World Internet Conference Wuzhen Summit opens. Alibaba Group CEO Wu Yongming said at the meeting that Alibaba is becoming an open technology platform company and will provide infrastructure for AI innovation and transformation in thousands of industries. The Wuzhen Summit of the World Internet Conference has entered its tenth year. AI technology represented by large models has developed rapidly and is profoundly reconstructing the digital world and the real world. The computing paradigm is undergoing fundamental changes, and the importance of AI computing is surpassing traditional computing and becoming the cornerstone of the digital world. Wu Yongming said that everyone has his or her own imagination of the world in the AI era. Some sound far-fetched, but some are becoming a reality.

Insilicon Intelligent launches AI-assisted decision-making automation laboratory: target discovery and verification completed in 14 days

Link: https://news.miracleplus.com/share_link/11532

AI pharmaceutical company Insilicon Intelligent released “the world’s first automated laboratory assisted by AI for decision-making.” According to reports, the “unmanned operation” laboratory covers an area of 1,600 square meters and integrates artificial intelligence with automation, robotics and biological capabilities. It can perform single tasks, such as high-throughput screening, high-content imaging, second-generation sequencing, etc. , it can also realize an integrated series process, such as a fully automated closed-loop dry and wet experiment that completes target discovery and verification within 14 days.

Shanghai releases 11 measures to promote the development of large AI models and support the settlement of relevant talents

Link: https://news.miracleplus.com/share_link/11533

Shanghai City recently issued a notice on the issuance of “Several Measures in Shanghai to Promote the Innovative Development of Large-Scale Artificial Intelligence Models (2023-2025)”, launching 11 measures to promote the innovative development of large-scale pre-training models and accelerate the creation of a world-class artificial intelligence industrial cluster.