The head of the artificial intelligence unit in Microsoft Mustafa Suleyman (Mustafa Suleyman) in an effort to strengthen the company’s position in the development of interactive Jeenets expands the Microsoft team, lures the key specialists from Google, including the creators of the Audio OverViews technology. Former Suleiman colleagues from Deepmind, from where he left in 2022, will be involved in the project to create multimodal models for processing text, sound and video.
Image source: Copilot
According to Financial Times, among the new Microsoft employees were Marco Tagliasacchi and Zalán Borsos, the creators of the Audio OverViews function, which allows you to transform the text in the style of a fascinating subcust. These researchers also participated in the development of Astra, the promising II-agent of DeepMind, capable of answering real-time questions using video, audio and text.
Matthias Minderer also joined the team. He will be engaged in the development of AI capabilities to analyze images. All three will work in the new Microsoft research center in Zurich and, according to a source familiar with the situation, will play a key role in the development of the next generation Copilot, on the basis of which interactive Jeenets capable of performing a wide range of tasks will be created.
Regarding personnel losses of this level, Google DeepMind did not give comments. However, it is noted that the transition of these specialists from Google DeepMind to Microsoft is part of a fierce struggle for Talents in the field of AI. At the same time, both companies remain key players in the development of multimodal AI models that are able to analyze and understand content based on audio, videos or images.
Against this competitive race, other companies are also not far behind – Openai introduced the voice mode for Chatgpt, Amazon introduces AI into its voice assistant Alexa. Google is also preparing for the release of Astra voice agent in 2025.