AI Doesn't Care Where You Work From
In a recent Stanford Biz School lecture, Eric Schmidt blames work at home for Google losing the AI battle to OpenAI and Anthropic (and in my mind, Meta/Llama3).
I think he is just deflecting. Here are my reasoning why Google is losing the AI battle:
1. Pioneer's dilemma - being first will almost never make you the eventual winner; someone younger, smaller, more nimble and hungrier will beat you at the game you invented
2. Google really isn't a pioneer on AI (or a lot of innovations people are falsely giving it credit for) - Almost all innovations at Google are via acquisitions; very few are home-grown. And we all know acquired talent don't last long in acquirer company, or work as hard/motivated.
3. Google's Search and Ad revenue may be over 80% of it's total revenue, but that's not where their engineers and scientists focus is on. I'm willing to bet that less than 2% of Google engineers and scientists actually are experts in search, big data and AI. The "man behind the curtains" is running 80%+ of the revenue show. Rest of engineers do everything else Google does, and most of Search and Advertising team are sales and legal - typical pattern of a mature product and industry.
4. Legacy Revenue Dilemma - When 80% of your revenue is on search and advertising, you're not going to kill the goose that's laying your golden eggs. Google doesn't know how to monetize on AI yet (nobody does), so they cannot shift their focus and resources to AI fast enough; their shareholders won't let them.
5. AI is a very human-centric thing. Google is Engineering driven, not Product/UX/UED driven. It got lucky with Search because it was pioneer and defined Search as returning a bunch of title/description/url listings and let people sort it out. ChatGPT is anti-Search in UX because it allowed natural language input and output, and future multi-modal AI will continue to mold input and output to what average humans want. No way Google Engineers will ever figure this out. If I was Eric Schmidt or Sundar - I would focus Google on solving engineering problems, that's what it is good at, not average human problems.
<Work From Home AI photo generated by Meta AI Llama3>