Microsoft’s AI boss, Mustafa Suleyman, has confirmed that the company is developing its own “off-frontier” AI models, which are 3 to 6 months behind OpenAI’s cutting-edge frontier models. This move comes as Microsoft seeks to build its AI capabilities in-house and reduce its dependence on OpenAI.
According to Suleyman, building off-frontier models allows Microsoft to focus on specific use cases while reducing development costs. He stated that this approach is part of the company’s strategy to play a “tight second” to OpenAI, given the high capital intensity of these models.
Microsoft has raised concerns about OpenAI’s GPT-4 model being too expensive and not fast enough for consumer needs. The company has also lost its exclusive cloud provider status to SoftBank, which led to a significant funding round for OpenAI that pushed the ChatGPT maker’s market cap to $300 billion.
Despite this, Microsoft is investing heavily in its AI capabilities, with recent announcements including Copilot Vision, Deep Research, Pages, Copilot Avatar, and Memory. The company aims to build self-sufficient AI capabilities internally but acknowledges that cutting-edge models may not be necessary for its use cases.
As Microsoft continues to scale its AI ambitions, it will be interesting to see how the company mitigates emerging issues with AI as it demands more computing power, GPUs, and cooling water.
Source: https://www.windowscentral.com/software-apps/mustafa-suleyman-confirms-off-frontier-ai-models-behind-openai