If you haven't been clued in yet, Google just dropped their latest bombshell - the preview of their new model, Gemini. I’m surprised to say this but it's nothing short of mind-blowing even if they cherry picked. Remember Bard? Yeah, that attempt that even Google's own engineers thought was a miss - not as disastrous as Tensorflow, but not exactly a home run, more like a drowning cat. Google was floundering, lagging in the AI race so visibly it hurt, they were in trouble. They even hit the panic button internally with a ‘code red’ last December, tossing out all plans for 2023 to pivot in response to ChatGPT's game-changing moves. Google's search engine was starting to feel like a relic from 2001, and who wants to step back in time when you've tasted 2023? Time is too previous to go back in time spending hours searching through the landfill of knowledge the old way.
Now, is Google's Sundar Pichai the new Michael Jordan of tech CEOs? Not yet…
Satya Nadella has been running laps around Sundar for years. Sundar might've lucked out here, but the real MVP? That's Demis Hassabis from OpenAI. Sure, Sundar gave Demis the resources to duke it out with OpenAI, but that's CEO 101, even Bran Blight the CFO/CLO turned CEO can do that. The real genius in AI? That's all Demis.
Remember when Demis boldly claimed back in June that Google DeepMind's next algo would outshine ChatGPT? I seemed like a reach for attention, jealous maybe "Hey, remember us?" from the team that freaked out China with their AlphaGo win. Well, guess what? Demis and his crew came through, big time.
But here's the kicker: this is now a "Code Red" for Microsoft and OpenAI. They're in deep trouble, and I'll break it down for you.
Why OpenAI (and by extension, Microsoft) should be sweating bullets:
Some might say I'm jumping the gun because Gemini Ultra isn't out in the wild yet. But Google isn't one for smoke and mirrors; they're the tortoise to OpenAI's hare, showing their hand only when they're good and ready. They're not fans of the spotlight, especially the negative kind (think Timnit Gebru, Maven project, that whole sentient AI fiasco...). The real threat? Google's mastery in AI chips with their TPU. While Gemini Ultra might be burning cash right now, I bet Google's racing to get an inference-only chip out. OpenAI, with all its drama (looking at you, Sam's AI chip side hustle in the news with its own turmoil), can't keep up. Google's a well-oiled machine, sans the drama. Sam can’t save them, they’re going to need Satya.
Now, about Google's Achilles' heel - their product development, specifically user experience and customer empathy. They've got the tech chops, sure, but their AI products and cloud offerings? Don’t get me started on Tensorflow. A mixed bag. Take Google Maps and Gmail - top-notch. But their AI product flops and the cloud confusion? What do you mean? Home Depot loves Google cloud… It's not about usability; it's about retailers like Home Depot hating Amazon. Retailers despise Amazon, and that's the only reason they're cozying up to Google's cloud. If Google drops the ball, it'll be on the product and go-to-market side, not the tech.
Over at Microsoft/OpenAI's camp, investors are probably biting their nails. OpenAI's revenue might not just dip; it could nosedive back to 2022 if Google launches a Gemini-Ultra-specific chip. Imagine OpenAI sliding from a billion-dollar revenue to scraping under $100M, unlikely but not impossible today. Microsoft's stock would feel the heat too. Their only move? Hustle out a Gemini-Ultra-rivaling demo, and fast. But realistically? They'll probably drum up some buzz around Q* to keep their user base from jumping ship to Google.
And what about Satya, the so-called Michael Jordan of tech CEOs? I bet he's lighting a fire under OpenAI, maybe even plotting Microsoft's direct involvement. If anyone can churn out a GPT4/GPT5 inference-only chip pronto, it's Satya, not Sam Altman. What about the equity upside in an inference-only chip company? Listen kids, the adults need to drive now. It's Microsoft's game to lose now and Satya isn’t about to lose.
This is a win for the AI community. A one-horse race isn't fun for anyone, and thankfully, the AI race just got a whole lot more interesting. It reminds me of the cutthroat semiconductor memory market battles of 2008-2012. You're either raking in billions or hemorrhaging them. Right now, as of yesterday, Microsoft's on the losing end, and they need to flip that script, pronto, all eyes on Satya now the game isn’t over yet.
Last thought, Satya also might decide he doesn’t have the talent in OpenAI to win and Microsoft could turn up their M&A like DeepMind did. As usual would love to hear everyone’s feedback/comments.
Thank you Jepson for the article. Apart from the well crafted summary on the topic, it was also your scripting that kept me locked in to read the full article. I loved it and appreciate this POV.
Cheers