If you believe Google's marketing hype, AI in a phone is really important, the best AI is the AI Google offers, and the best place to get that AI is on Google's flagship smartphone, the Pixel 8. We're five months away from the Pixel launch 8, and that doesn't seem like a justifiable position anymore: Google says its latest AI models can't run on the Pixel 8.
Google dropped this news in a Mobile World Congress closing video spotted by Mishal Rahman. At the end of the presentation in a Q&A session, Google employee Terence Zhang, a member of the Gemini-on-Android team, said, “[Gemini] The Nano will not be coming to the Pixel 8 due to some hardware limitations. It's currently available on the Pixel 8 Pro and recently available on the Samsung S24 family. “It will reach more advanced devices in the near future.”
This is a wild statement. Gemini is Google's latest AI model, and it was a huge hit when it launched last month. Gemini comes in several different sizes, with the smallest “Nano” being specifically designed to run on smartphones as a much-touted “on-device AI.” The Pixel 8 and Pixel 8 Pro are Google's flagship smartphones. Google designed the phone, the chip, and the AI model, and somehow it can't make these things work well together?
What makes it even stranger is that the Gemini Nano can work on the Pixel 8 Pro but not on the smaller Pixel 8 due to “hardware limitations.” What restrictions would those exactly be? The two phones have the same Google Tensor SoC. They run the same program. The main differences between the two phones are the screen size (6.7 inches vs. 6.2), battery size, different camera loading, and 8GB vs. 12GB of RAM. RAM is the only known difference you can point to that could lead to processing limitations, but the Gemini Nano also works on the Galaxy S24 series, where the base model has 8GB of RAM. That RAM is the issue means Samsung phones are more efficient at using RAM than Pixel phones, which is hard to believe. If the Pixel 8 Pro Tensor 3 and Pixel 8 Tensor 3 are different in some way, it's not in the spec sheet.
Five months ago at the Pixel 8 launch event, Google painted a picture A very different picture From the Pixel 8 series: “I'm thrilled to introduce you to the next evolution of AI at your fingertips, Google Pixel 8 Pro and Google Pixel 8. Our latest phones combine many technologies from across Google. They are the first phones of their kind to use our latest Google Tensor chipset. Featuring the best Android experience, first-of-its-kind camera experiences, and Google's latest AI advancements, both devices feature a custom Google Tensor 3 SoC Google claimed “It's specifically designed to bring Google's AI breakthroughs directly to Pixel users and show the world what's possible.” This custom, AI-focused design from Google was supposed to deliver “incredibly useful experiences that no other phone can deliver.”
When two phones are launched simultaneously, it is always difficult to distinguish the actual differences between the two models. Sometimes, the devices are talked about in the plural, while other times “Pixel 8” is used to represent both devices. Sometimes, the most expensive device is mentioned individually for no reason other than that it's the most expensive flagship device. Between the hour-long presentation and the special press briefing that Ars was a part of, “what's the difference” became a very trite question that was expected to be answered clearly. The specifier here is usually the spec sheet, which is expected to explain in plain language what you're actually buying. Google store has it Comparison page You can put the Pixel 8 and Pixel 8 Pro directly against each other, and nothing shows a difference in AI processing capabilities or a difference in Tensor chips.
In the case of the Pixel 8 and Pixel 8 Pro, Google wasn't clear enough in its communications at launch. Today, though, I'm rewatching Launch the offer With the new knowledge that is out there He is Some big difference in AI processing capabilities, you might pick up on some phrases like talking about “on-device LLM in the Pixel 8 Pro” which you can now interpret as an announcement of AI capabilities exclusive to the Pro model, but that wasn't clear at the time.
As a consumer, it's hard not to feel misled, which is embarrassing for Google, but to care about this in practice, you'll need to know what the Gemini Nano actually does and why you should care about it. This is a difficult question to answer. Google has a page up here Details of some of the features the Gemini Nano runs on the Pixel 8 Pro, but the feature can also be run by different models on different devices. However, the list lists Google Recorder's “Summary” feature and Gboard's “Smart Reply” feature. A lot of Google apps already have Smart Reply without the Gemini Nano. Third-party developers can also plug the Gemini Nano's embedded model into their apps, but it's hard to imagine anyone doing that with such limited hardware support.
The other option is to forget about doing all this AI stuff on the device and just do it in the cloud. As a great example of this, none of these Gemini Nano items have anything to do with Google's Gemini Chatbot, which all runs in the cloud. The big question is what this will mean for the smaller Google Pixel 8 in the future. Google promised seven years of OS updates for new Pixels, and that it's actually stripping features due to “hardware limitations” after five months is disappointing.