Microsoft’s OpenAI partnership was born from Google envy

It turns out that the current AI landscape is rooted in – what do you know – fear, jealousy and intense capitalist ambition. Emails revealed in the Justice Department’s antitrust case against Google, first reported by Business Insider, show Microsoft executives expressing concern and envy over Google’s lead in AI. This sparked an urgency that led the Windows maker to initially invest $1 billion in its now-indispensable partner, OpenAI.

In a strongly worded 2019 thread titled “Thoughts on OpenAI,” Microsoft CEO Satya Nadella relays a lengthy message from CTO Kevin Scott to CFO Amy Hood. “Very good email that explains why I want us to do this…and also why we will then ensure that our infra managers execute,” Nadella wrote.

Scott wrote that he was “very, very concerned” about the rapid growth of Google’s AI capabilities. He says he initially dismissed the company’s “gaming stunts,” likely referring to Google’s AlphaGo models. One of them beat Go Ke Jie world champion in 2017, a remarkable feat at the time. (Later Google models outperformed this one, completely removing the need for human training.)

But Scott says neglecting Google’s gaming progress “was a mistake.” “When they used all the infrastructure they had built to create (natural language) models that we couldn’t easily replicate, I started to take things more seriously,” Scott wrote. “And as I dug in to try to understand where all the capability gaps were between Google and us when it came to training models, I became very, very concerned.”

Scott talks about how Microsoft struggled to copy Google’s BERT-large, an AI model that deciphers the meaning and context of words in a sentence. Scott placed the blame on the infrastructural advances made by his rival – not Microsoft.

“It turns out that simply replicating BERT at scale was not easy for us to do. Even though we had the model for the model, it took us about 6 months to train the model because our infrastructure was not up to the task,” wrote the Microsoft CTO. “Google had BERT for at least six months before that, so in the time it took us to build the ability to train a 340 million parameter model, they had a year to figure out how to put it into production and move on to larger scale and more interesting models.

He also admired and envied the autocomplete capabilities of Google’s Gmail, saying they were “getting awfully good.” He commented that Microsoft was “several years behind the competition in terms of (machine learning) scale.” He commented on the “interesting” growth of OpenAI, DeepMind and Google Brain.

Scott touted Microsoft’s “very smart” people on its machine learning teams, but said their ambitions were being held back. “But the deep learning core teams within each of these larger teams are very small, and their ambitions have also been limited, meaning that even if we start to provide them with resources, they still have to go through a learning process to grow,” Scott wrote. “And we are several years behind the competition in terms of ML scale.”

After explaining to Hood that Scott’s concerns were “why I want us to do this” — investing in OpenAI — the company accommodated its CEO’s wishes. Microsoft invested $1 billion in the Sam Altman-led startup in 2019, and the rest is rapidly evolving history. (It has now invested $13 billion.) It’s a technology that does incredible things but threatens to gut the job market and give propagandists their most powerful tools yet in what was already an era widespread disinformation.