Google and Microsoft’s competition over technology’s AI future continues to intensify. Google may have turned to a company started by former OpenAI employees, despite Microsoft’s close ties to ChatGPT creator OpenAI: the Anthropic, a little-known group.
Google invested approximately $300 million in the startup in late 2022, though the news was not reported at the time. Anthropic is required to purchase cloud computing resources from the search engine giant in exchange for the money. Google received a ten percent stake in the company.
This dynamic is somewhat comparable to Microsoft’s partnership with OpenAI. OpenAI provides the research expertise, and Microsoft provides access to its enormous cloud platform, which is required to train the most recent compute-intensive AI models, in addition to investing billions of dollars.
It is noteworthy that Anthropic is also working on its own chatbot with a variety of uses: Claude, a potential rival to ChatGPT. A comparison of the systems of Anthropic and OpenAI can be found here, even though Claude is currently in closed beta.) However, it is unclear whether Google intends to incorporate Claude into its services in the same manner that Microsoft does with ChatGPT. One of the reasons for the deal is simply expanding Google’s cloud computing business. Google already has a lot of in-house expertise developing AI language systems.
The circumstances surrounding Anthropic’s beginnings are also relevant. Dario Amodei, who previously held the position of vice president of research at OpenAI, established the organization in 2021 as a public benefit corporation. Amodei brought along a number of OpenAI researchers, including lead engineer Tom Brown for the AI language model GPT-3. Amodei left OpenAI “after a disagreement over the company’s direction,” specifically the company’s shift toward a more commercial focus following its first deal with Microsoft in 2019.
Since then, a lot of AI researchers have said that OpenAI’s actions have become more reckless, especially when it put ChatGPT on the public web late last year without using software or proper safeguards to reliably detect its output. On the other hand, Anthropic places a strong emphasis on the development of “reliable, interpretable, and steerable AI systems” on its website. But will these priorities change as a result of Google’s investment?
In any case, we can anticipate learning more about Google’s AI plans in the near future. On February 8th, the company will hold an event on the subject. Additionally, Microsoft’s plans to integrate ChatGPT into Bing are likely to be discussed in greater detail in the near future.