In a recent blog post, Bill Gates predicted the artificial intelligence (AI) boom of 2023 is set to accelerate in the next 12 months, driving further innovation across education, medicine and healthcare.
That view is clearly shared by Microsoft, the company Gates founded and still acts as an advisor for. Throughout the past year, the firm has made AI the driving force of its product development strategy, injecting machine learning (ML) capabilities across its software offering.
Just two years ago, Microsoft’s role in AI development was largely overshadowed by Google. Things have changed since then.
Kickstarted by the launch of ChatGPT in November 2022, generative AI has emerged as a fiercely competitive technology sector in 2023. Developers like OpenAI and Anthropic are functioning as proxies for the Big Tech heavyweights that have funded their growth.
Microsoft’s partnership with OpenAI has allowed it to integrate the latest GPT large language models (LLMs) with products such as Bing Search, MS Office and Azure Cloud. Meanwhile, OpenAI’s DALL-E models now power products such as Microsoft Designer and Bing Image Creator.
Independent of OpenAI’s GPT and DALL-E family of ML programs, Microsoft has also continued to develop its own AI.
At the beginning of the year, it launched VALL-E, a new AI tool for text-to-speech synthesis (TTS).
More recently, the firm has debuted a suite of small language models (SLMs) called Phi . These are part of a burgeoning field of SLMs. These promise to deliver LLM-style language and reasoning capabilities for a fraction of the cost.
While LLMs are some of the AI programs aiming to improve efficiency, the computational burden of training and inference continues to limit progress.
Through its cloud business, Azure, Microsoft is one of a handful of “hyperscalers”, providing much of the muscle behind AI innovation so far. As AI services have become an increasingly integral source of revenue for Azure, Microsoft has sought to optimize its data centers, installing more GPUs and even unveiling its first custom AI chips in November.
Looking ahead to 2024, it seems likely that Microsoft will continue its strategy of integrating OpenAI’s powerful AI models into its software.
Windows 12 is expected to come out in 2024. There has been speculation Microsoft could embed chatbot functionality directly into its operating system. The firm has yet to make any announcement on the matter. However, Windows’ virtual assistant Cortana was discontinued this year. Therefore, some kind of GPT-powered alternative would serve as a logical replacement.