Amar Subramanya will likely become one of the most closely watched hires Apple has made in years. Here’s what he needs to do first.
Apple’s new VP of AI has some experience with catching up
Following the retirement announcement of John Giannandrea, Apple confirmed that Amar Subramanya will join the company after less than six months as Corporate Vice President of AI at the newly formed Microsoft AI division.
Before that, though, he spent 16 years at Google, where he joined as a Staff Research Scientist and became a Principal Engineer before being appointed Vice President of Engineering in 2019.
He was also credited in two recent papers of very high-profile releases: Gemini in December 2023, and Imagen 3 in August 2024.
Interestingly, Google spent a good portion of the last few years in a similar predicament as Apple.
Like the rest of the market, Google missed the starting gun with the release of ChatGPT, and scrambled before finding its footing, which was particularly awkward for them, since Google invented the Transformer architecture (the T in GPT).
But as anyone who follows the AI field closely will tell you, Google is undeniably back, and it is giving OpenAI a run for its money. Literally.
Following the release of the most recent Gemini model, OpenAI’s CEO Sam Altman told his team to expect some “rough vibes” and “temporary economic headwinds,” which translates to: “they caught up with us, this is not going away, and investors know it, too.”
Google’s path to not-AGI-but-prett-mych-the-same-intelligence-as-OpenAI was not without its… detours. Remember when Bard was the company’s bet on the future of AI? Good times.
The thing is: it took them three years (almost to the day since the release of ChatGPT) to deliver on a mid-flight course-correction, but they did it.
And while it’s too early to tell whether they’ll permanently beat OpenAI, that is not even the point. The point is that not only do they finally have an undeniable state-of-the-art model, but they also have every single Google-owned platform to deploy it on, which is something that nobody else has.
And while Subramanya left Google in June, he was obviously closely involved in the bulk of the development of these new Gemini models, being the company’s VP of Engineering.
But what about Apple?
From day one, Subramanya will obviously face immense pressure to deliver on the perceived fumbles of his predecessor.
On the bright side, that buys him some goodwill and, to some extent, some time to, first, get things in order, and second, to get them right.
Still, he’s got an uphill battle ahead of him, with a heavily shorthanded team at an all-time low morale.
And while it’s impossible to know for sure from the outside, it’s reasonable to assume that he will need to bring in reinforcements rather quickly, if he hopes to recover from the relentless waves of losses Apple has suffered, particularly over the past year.
The good news is that with new leadership at the helm and all the makings of a somewhat fresh start, Apple might have a chance of attracting talent that, until earlier today, saw little to no future in a career in machine learning at the company.
Some of them, very possibly, will be ex-colleagues of Subramanya.
And then, there’s the outside expectation, which isn’t exactly his responsibility, to a certain extent. He’s there to fix Apple’s underlying AI technology, and it will be up to marketing and product to choose how to tell this story.
But that story will rely exclusively on what he actually manages to deliver, by carefully threading the needle between acting quickly to show what he came for, while making sure that Apple can actually deliver on its AI vision, even if it means starting some portion of its work from scratch.
And finally, there’s the long-term vision. While rebuilding confidence (both within and outside the company) will be the bulk of his work, this will still be just the beginning.
For the past few years, we’ve grown accustomed to thinking that Apple’s role in machine learning is simply to catch up to the competition. But that’s not it at all.
The job of any company that wishes to make meaningful advancements in AI is to foster an environment where research and application can reinforce each other, which, ironically, was what John Giannandrea set out to do before ChatGPT turned the industry upside down.
And while it feels unfair that, despite contributing far more than people will ever give him credit for, Giannandrea will likely be remembered as the guy who dropped the ball on AI at Apple, his tenure should also serve as a cautionary tale to Subramanya: fixing what’s broken today is only part of the job. Just as important will be laying the foundations for the next decades of a technology that won’t ever pause, slow down, or be finished.
What’s your take on Apple’s AI leadership change? Let us know in the comments.
Great Black Friday 2025 deals
FTC: We use income earning auto affiliate links. More.
