In 2017, Google was criticized for a bad emoji. This week, CEO Sundar Pichai fixed it. How he did this says a lot about the progress Google has made in generative AI.
The internet giant schooled the rest of the tech industry when it launched its latest model, Gemini 3, on Tuesday. It followed that up by rolling out an upgraded image-generation tool called Nano Banana Pro, which is churning out impressive, realistic pictures, diagrams, and charts.
Google proved that AI scaling laws still work, just as its rivals are being questioned. The stock jumped to a record, making the company worth more than Microsoft.
This is all a long way from 2017. Back then, Google rolled out a cheeseburger emoji for Android smartphones. It had the cheese under the meat. Absolute scandal.
I think we need to have a discussion about how Google’s burger emoji is placing the cheese underneath the burger, while Apple puts it on top pic.twitter.com/PgXmCkY3Yc
— Thomas Baekdal (@baekdal) October 28, 2017
Pichai apologized and said Google would get working on a fix immediately, in a joking kind of way.
Fast forward to this week with the launch of Gemini 3 and Nano Banana Pro.
These new AI models are much better at creating and rendering images. Their understanding of 3D space and how the world works, physics-wise, has improved so much that they now absolutely nail the cheeseburger stack.
Pichai tweeted an AI-generated image out to prove the point, writing “iykyk” as a nod to the 2017 furor. As you can see, the earth has been set back on its correct axis, and we can all calm down: the cheese is above the meat.
OK, it’s time to get serious. Why am I telling you this?
The ability of AI models to innately understand where stuff should go in the world is really important.
“Normally, AI models struggle with spatial orientation, particularly with respect to the relative position of objects,” Balaji Srinivasan, a tech investor and former Coinbase CTO, wrote on X after Pichai’s latest burger post. “But this image (if rendered by Gemini 3) seems to resolve that issue, as the exact spatial positioning of the cheese is handled correctly and precisely.”
If AI models can know where cheese should go in a burger, they might also know where more important stuff should be in the real world. That could mean better machine decision-making in design, engineering, and other fields.
One theoretical example: A safety barrier likely needs to be placed in the right spot on the corner of a road. Maybe AI models can guide workers to put this structure in exactly the best spot, down to the millimeter.
There’s another takeaway from this eight-year cheeseburger saga. Google has been criticized for being behind in generative AI, and this week’s releases have finally put those questions to bed.
Powerful products like Gemini 3 take many years, and a lot of technical research and plumbing, to pull off. Google has been working at this for a very long time, and Pichai has been pushing the company toward an AI-first mindset for about a decade. Now the fruits of these labors are showing through.
“Google really did drop everything they were doing to truly focus on AI. And Gemini 3 represents the moment when they actually retook the lead, at least for now,” Srinivasan wrote. “When combined with Sundar doubling Google’s revenue to $100B, he’s proven he can lead Google to unprecedented heights both technologically and commercially.”
“Hence: iykyk. If you know, you know,” he added.
Sign up for BI’s Tech Memo newsletter here. Reach out to me via email at [email protected].
