MIcrosoft tops CTO Kevin Scott, CEO Satya Nadella and co-founder Bill Gates were very afraid in 2019 that Google was about to overtake them in the AI segment. The story should end with Microsoft integrating and launching AI only after three years of heavy investment in OpenAI.
Gates potentially more active in Microsoft than thought
It reveals internal emails dubbed “Thoughts about OpenAI” between the summits in 2019. You may then wonder why we know this: the reason is the US Department of Justice’s case against Google. Scott stated that the company was several years behind Google in terms of scaling AI, and therefore already in 2019 ended up investing $1 billion in OpenAI – since then they have injected another $12 billion.
The company found out in the summer of the same year that it was the scaling that was at issue. He explains that it took Microsoft half a year to copy Google’s LLM Bert and to train it because Microsoft’s hardware park was not powerful enough. Scott only discovered this when he took a deep dive into the LLM technology and found that Google was way ahead of them in terms of training the models: “I got very, very anxious,” the newly revealed emails reveal also reveals a specific feature that impressed at the time: Gmail’s autocomplete capabilities.
As for Bill Gates, there is little news in the media at the time of writing about his work at Microsoft, but at least in 2019 and ever since 2016, Gates must have been important in getting an agreement with OpenAI. In relation to longer strategic roadmaps, it is not impossible that he is still an important player.
