Tencent Acquires Techland, the Developer of Zombie Games, Cohere Introduces AI Assistant & Stability AI Reveals New Language Models
Hello! 🙋We have the following updates for you today! 📣
- Zombie game maker Techland was acquired by Tencent’s gaming giant
- Cohere releases Coral, an AI assistant designed for businesses
- Stability AI unveils new FreeWilly language models
Zombie Game Maker Techland Was Acquired by Gaming Giant Tencent
According to the Techland blog, Tencent, which is a major gaming company, is in the process of acquiring the majority of its shares. The blog also states that they have partnered with one of the finest video game companies that has helped many to reach new heights.
Techland is known for open-world zombie games like Dying Light. The specific terms of this transaction are not disclosed.
Among the largest gaming firms globally, Tencent has extensively invested in numerous studios and businesses worldwide. They acquired Riot Games, the creator of League of Legends, and possess stakes in Krafton (the force behind PUBG), Supercell (the developers of Clash of Clans), Ubisoft, and Epic Games.
Cohere Releases Coral, an AI Assistant Designed for Businesses
Cohere announced AI-powered knowledge assistance, Coral, to help businesses enhance the productivity of diverse teams like finance, support, sales, and others. Coral enables firms to optimize capabilities and access relevant data sources, enriching their knowledge base.
It is built on Cohere’s, a high-performing generative language model that undergoes weekly retraining. Cohere offers natural language processing (NLP) solutions powered by large language models (LLMs) and recently secured $270 million in a funding round.
Prominent industry leaders, including Nvidia, Oracle, and Salesforce Ventures, led this investment. Consequently, the company’s valuation has exceeded $2 billion.
Stability AI Unveils New Freewilly Language Models
FreeWilly1 and FreeWilly2 are two strong new open-access large language models (LLMs) that Stability AI has developed in partnership with CarperAI lab. These models demonstrate extraordinary reasoning abilities across a range of criteria.
Using a synthetically created dataset and the LLaMA 65B foundation model, FreeWilly1 was improved. The LLaMA 2 70B foundation model was also used to build FreeWilly2, which for some jobs, achieved performance similar to GPT-3.5.
In the area of open-access large language models, FreeWilly1 and FreeWilly2 are ground-breaking models. The models are anticipated to open up many possibilities and motivate fresh AI applications.
Both models have displayed exceptional performance in reasoning, linguistic subtleties, and problem-solving in Law and mathematics.
For more updates, check out the Latest Tech category.
