(Credit: Tada Images/Shutterstock.com)
Ignoring Google's AI efforts was a "mistake," a Microsoft exec said in a 2019 email exchange sent shortly before Redmond's billion-dollar investment in OpenAI.
Microsoft is "multiple years behind the competition"—like Google—when it comes to AI and machine learning, Microsoft CTO Kevin Scott wrote to Bill Gates and CEO Satya Nadella.
"I was highly dismissive of [Google's] efforts. That was a mistake," he wrote. "When they took all of the infrastructure that they had built to build [natural language processing] models that we couldn't easily replicate, I started taking things more seriously. And as I dug in to try to understand where all of the capability gaps were between Google and us for model training, I got very, very worried."
The internal emails, which include large blocks of redacted content, were made public this week as part of the ongoing US Department of Justice antitrust lawsuit against Google. The New York Times previously reported that Google had tried to keep documents pertaining to the trial a secret, but the judge in the case ruled last year that the DOJ could publicize some case files. Last week, media outlets pushed for the release of the 2019 emails, and the judge allowed it because they "shed light on Google's defense concerning relative investments by Google and Microsoft in search," Bloomberg reports.
In one email, Scott noted that OpenAI and Google had large-scale ambitions, but Microsoft struggled to make AI advancements with the BERT deep-learning model because of infrastructure constraints while Google succeeded.
"Google had BERT for at least six months prior to that, so in the time that it took us to hack together the capability to train a 340M parameter model, they had a year to figure out how to get it into production and move on to larger scale, more interesting models," Scott continued, adding that Gmail's auto-complete email abilities were "scarily good."
The CTO also said Microsoft's deep learning teams working on AI were "very small" and that Microsoft is "multiple years behind the competition" because of its lack of investment and resources into AI research and development.
In response, Nadella wrote: "Very good email that explains why I want us to do this. And also why we will then ensure our infra[structure] folks execute."
A month later, Microsoft announced a billion-dollar investment in OpenAI. Last year, it made another "multi-billion-dollar" investment in the company reportedly worth $10 billion.
Since those emails, Microsoft has dramatically ramped up its AI development and product launches, mostly notably integrating its Copilot AI assistant across Windows and developing specialized versions of the model, like Copilot Designer. It also released GitHub Copilot with help from OpenAI, which helps software engineers write code, and this week pushed out a dedicated GitHub Copilot Workspace for developers.
Google, meanwhile, rebranded its Bard chatbot to Gemini and launched a smaller AI model, Gemma, inspired by Gemini. It's also launched two new coding AI tools, CodeAssist and CodeGemma, designed to speed up software development.
Expect more AI news from their Build and I/O developer conferences later this month. Apple, meanwhile, is also gearing up for a big AI reveal at its Worldwide Developers Conference in June.
One thing Microsoft probably won't mention: The New York Times lawsuit that accuses the company and OpenAI of scraping the newspaper's content without permission to feed their AI models. Microsoft has written off the claims as "doomsday futurology," but it's now facing a similar lawsuit from eight papers owned by Alden Global Capital, including The Chicago Tribune, The San Jose Mercury News, and The Denver Post.