- Microsoft to sell a new version of Databricks’ software via Azure.
- OpenAI’s ChatGPT has had its challenges with enterprise adoption.
- Microsoft’s partnership dynamics shift towards other AI providers.
Microsoft, after collaborating with OpenAI on projects like ChatGPT, is now setting its sights on Databricks. The tech company intends to offer a new variant of Databricks’ software through its Azure cloud-server unit. This move will enable businesses to develop AI applications, either from the ground up or by adapting open-source models, as an alternative to OpenAI’s proprietary versions.
Microsoft’s Relationship with OpenAI
Major tech enterprises such as Apple, Spotify, Wells Fargo, Samsung, JP Morgan, and Verizon previously moved away from ChatGPT, restricting its use among their staff. Despite OpenAI’s endeavors, it seems to have faced challenges in gaining the full confidence of the enterprise sector.
Interestingly, while offering the Azure-Databricks service, Microsoft is utilizing OpenAI’s technology to create a chatbot reminiscent of ChatGPT. This tool aims to help users with limited tech proficiency navigate Databricks’ software. Consequently, some Microsoft clientele may opt for open-source solutions over OpenAI’s closed-source ones.
Microsoft had formerly introduced “Azure ChatGPT” on GitHub, promoting it as a secure enterprise solution, noting data would remain protected on Azure and would not be shared with OpenAI. Additionally, Microsoft engaged with IBM Consulting, which aims to assist businesses in deploying Azure OpenAI Service. This alignment followed IBM’s partnership with Meta, targeting the integration of Llama 2 into watsonx.ai, a potential rival to OpenAI’s GPT-4.
OpenAI has responded to certain concerns by clarifying their policies. They have confirmed that if users turn off the chat history functionality, chat histories aren’t used for model training. Sam Altman, from OpenAI, has mentioned that API-submitted data isn’t used for model enhancement unless users give clear consent.
Broader Partnership Strategies
Microsoft’s association with Meta has shifted its relationship dynamics with OpenAI. Aligning with Meta and utilizing Llama 2 has enabled Microsoft to diversify its partnerships. In a similar vein, offering Databricks services on Azure aligns with Microsoft’s strategy to compete with Amazon Web Services (AWS). AWS provides a range of models like AI21, Cohere, Anthropic Claude 2, and Stability AI SDXL 1.0, and Microsoft’s collaboration with Databricks could be seen as a move to provide diverse options to its clients.
Databricks has recently procured Mosaic ML, which developed MPT-30B, an open-source base model. It remains to be seen if this will be integrated into Azure in the foreseeable future.