Collection of Big Model Daily News on November 21st
[Collection of Big Model Daily News on November 21st] OpenAI is actively discussing “reunion”; more than 100 customers of OpenAI have contacted competitor Anthropic; OnePlus has launched an AI music studio, allowing users to freely create music and videos
JetBrains releases 2023 developer report: 77% of developers use ChatGPT chat assistant
Link: https://news.miracleplus.com/share_link/11873
JetBrains recently announced the “2023 Developer Ecosystem Status” survey report, which gathered survey results from 26,348 developers around the world. The survey results on the generative A service show that 77% of developers use the ChatGPT chat assistant, and 46% Developers use the GitHubCopilot programming assistant. The most common way developers use AI assistants is to ask general questions about software development using natural language. But 59% of developers have security concerns about using generative AI services.
Another AI company is in trouble! C3.ai reports massive layoffs, stock price plummets
Link: https://news.miracleplus.com/share_link/11874
Some media broke out that the artificial intelligence software company C3.ai implemented employee reductions last week, affecting multiple departments; the reasons given by the company were poor employee performance and the company’s need to save costs; once the news was made public, the company The stock price fell off a cliff, falling below $30.
OnePlus launches AI music studio, allowing users to freely create music and videos
Link: https://news.miracleplus.com/share_link/11875
OnePlus has launched an AI Music Studio on its overseas official website, allowing users to create and share AI-generated music.
Use retrieval to enhance generation to make large models more powerful. Here is a step-by-step Python implementation
Link: https://news.miracleplus.com/share_link/11876
Ever since it was recognized that large language models (LLMs) could be made more powerful using their own proprietary data, people have been discussing how to effectively integrate general knowledge of LLMs with proprietary data. People have been debating about this: Which is more suitable, fine-tuning or retrieval-augmented generation (RAG)? This article will first focus on the concept and theory of RAG. It will then be shown how a simple RAG can be implemented using LangChain for orchestration, the OpenAI language model, and the Weaviate vector database.
The GPT-4V Medical Licensing Examination score exceeds that of most medical students. How far will it take AI to join clinical practice?
Link: https://news.miracleplus.com/share_link/11877
The application of artificial intelligence (AI) in medical imaging diagnosis has made great progress. However, doctors often have difficulty accepting AI’s diagnostic results without rigorous testing. For them, understanding the judgments given by artificial intelligence based on medical images requires additional cognitive costs. In order to enhance the trust among doctors in artificial intelligence for assisting medical treatment, it may be an effective method to let the AI pass a qualification examination that doctors must pass. The Medical Licensing Examination is a standardized test used to evaluate a doctor’s professional knowledge and skills. It is the basis for measuring a doctor’s ability to care for patients safely and effectively. In the latest research, an interdisciplinary team of researchers from the University of Massachusetts and Fudan University used the generative multi-modal pre-training model GPT-4V (ision) to break through the best level of artificial intelligence in medical question answering. The study tested GPT-4V’s ability to answer questions on the U.S. Medical Licensing Examination, specifically those that contain images — something that has long been a challenge for medical AI systems. This study shows that GPT-4V not only surpasses its predecessors such as GPT-4 and ChatGPT, but also surpasses most medical students, providing a theoretical possibility for artificial intelligence to be used as a tool to assist diagnosis and clinical decision-making. This study analyzes the performance of GPT-4V in different medical subfields.
Replacing civil servants with AI? UK government plans to set up a task force
Link: https://news.miracleplus.com/share_link/11878
An artificial intelligence “task force” is about to be established at the center of Whitehall; its mandate will be to reduce the size of the UK civil service and increase public sector productivity. Deputy Prime Minister Oliver Dowden plans to form a task force of 30 “high-end, technically capable” experts in artificial intelligence and data engineering, with an annual budget of about £5 million, to jump-start public services transformation process.
OpenAI progress collection
Altman says top priority with Microsoft CEO is ensuring OpenAI thrives
Link: https://news.miracleplus.com/share_link/11879
Sam Altman, who was dismissed as CEO by the OpenAI board of directors, said on social media that Satya and my top priority remains to ensure that OpenAI continues to “thrive.” We will abide by our commitment to “provide comprehensive operational continuity to our partners and customers.” The partnership between OpenAI and Microsoft makes this very feasible.
OpenAI is actively discussing a “reunion”
Link: https://news.miracleplus.com/share_link/11880
OpenAI is in active discussions to “reunite” the company, according to a memo to employees on Monday from Anna Makanju, OpenAI’s vice president of global affairs.
OpenAI’s board of directors contacted rival Anthropic to discuss a merger
Link: https://news.miracleplus.com/share_link/11881
According to media reports citing people familiar with the matter, after firing Sam Altman, the OpenAI board of directors discussed the merger of the two companies with its main competitor, Dario Amodei, co-founder and CEO of star AI startup Anthropic. People familiar with the matter said that the OpenAI board of directors was trying to persuade Dario Amodei to take over the position of OpenAI CEO, so the merger was discussed. However, the person did not disclose whether there were “serious discussions” on the merger proposal.
Ultraman may still return to OpenAI, Microsoft CEO says: open to this
Link: https://news.miracleplus.com/share_link/11882
Microsoft CEO Satya Nadella recently stated that he is open to Sam Altman returning to OpenAI; Nadella said that whether Altman joins Microsoft or returns to OpenAI, the entrepreneur will Will cooperate with Microsoft; he also said that no matter what, OpenAI needs to reform its governance.
OpenAI over 100 customers contact competitor Anthropic
Link: https://news.miracleplus.com/share_link/11883
According to relevant media reports, the firing of Sam Altman, the former CEO of OpenAI, continues to ferment, and doubts about the future of OpenAI continue to rise. Meanwhile, OpenAI’s customers are looking for exits, suggesting there could be significant business losses that could hit the startup hard. It is reported that more than 100 OpenAI customers contacted OpenAI competitor Anthropic over the weekend. Cohere, another rival startup to Google Cloud and OpenAI, has also been contacted, people familiar with the matter said.