Ad
Skip to content
Read full article about: Alibaba's cloud computing unit debuts two open-source AI language models

Alibaba's cloud computing unit has released two open-source AI models, Qwen-7B and Qwen-7B-Chat, marking the first time a major Chinese tech company has open-sourced its large language models. Aimed at helping small and medium-sized enterprises leverage AI, the models' code, weights, and documentation will be freely available to the global research community and commercial institutions. However, companies with more than 100 million monthly active users will need a license from Alibaba to use the models.

Alibaba launched its closed-source GPT-4 competitor Tongyi Qianwen in April and its DALL-E and Midjourney competitor Tongyi Wanxiang in July.

Read full article about: ChatGPT update brings useful feature known from Microsoft's Bing Chat

ChatGPT now provides contextual questions and answers in each chat, as well as suggestions for starting a new chat on various topics, such as "Explain airplane turbulence". The questions and answers are displayed above the chat box and relate to the content already generated. This is similar to the suggestions that Microsoft displays in the Bing chatbot.

No official information about the new feature has been released by OpenAI yet.

ChatGPT generates contextual questions. | Image: THE DECODER
Read full article about: DeepL launches Europe's Largest NVIDIA DGX SuperPOD for advanced AI translation

Germany-based DeepL has installed an NVIDIA DGX SuperPOD to expand its LLM capabilities. Located in a Swedish data center, it is the "first commercial deployment of this scale in Europe," according to the company. Benchmarking showed a performance of 21.85 PFlop/s, which would rank it 26th in the world and eighth in Europe. The NVIDIA DGX SuperPOD, consisting of 68 NVIDIA DGX H100 systems, will help DeepL to train large language models faster and develop new AI communication tools for the global market.

Read full article about: BTLM-3B-8k-base brings LLM capabilities to devices with just 3GB of memory

Cerebras and Opentensor have trained a powerful 3 billion parameter language model with an 8k context length window, called BTLM-3B-8k-base, on the Condor Galaxy 1 (CG-1) supercomputer. This new model outperforms similar models, achieves performance comparable to open 7B parameter models, can be quantized to fit on devices with as little as 3 GB of memory, and is licensed for commercial use. It requires 71% fewer training FLOPs and has a 58% smaller memory footprint for inference than comparable 7B models.

Read full article about: Warren Buffett still cautious about investing in AI, despite being amazed by ChatGPT

Warren Buffett expressed cautious fascination with artificial intelligence and ChatGPT in a CNBC interview. While admitting that he doesn't understand AI enough to invest in it, the legendary investor still sees its potential for everyday use. However, he noted that he's interested in understanding how the technology evolves into a tangible, valuable business. Buffett also acknowledged the risks associated with AI, comparing it to the invention of the atomic bomb and expressing uncertainty about its long-term benefits to humanity.