|
|
|
|
|
by rurp
798 days ago
|
|
I recently got a new Pixel phone and Google's much hyped new AI features just seem so... gimmicky. One of the setup examples shows how you can circle a tent on the left side of a picture and move it more to the center. Neat, I guess? It's kind of a fun toy but I'm not sure what problem this is actually solving. I'm sure there are some usecases for this out there, but it's not a capability I have ever found myself wishing for. Meanwhile the rest of the phone is surprisingly buggy and annoying. Basic functionality I use every day is worse than on any other recent phone. Google has never been a great product org, but this desperate need to be seen as one of the cool kids in AI is making things worse. Granted I think of phones/computers more as a tool than a toy and put much higher value on usability and reliability versus novelty; perhaps I'm outlier in that. |
|
Not just that, but their biggest crime is that almost none of those fancy AI features Google paraded at the Pixel launch even actually run on-device but need to be sent to their cloud for processing, despite all the gloating about their new Tensor 3 chip's AI capabilities being the most important (since that chip sucks at CPU and GPU benchmarks compared to Apple and Qualcomm). Also, their Tensor 3 can't even run Google's smallest LLM. Absolutely embarrassing.
They REALLY need to unify the HW and SW development efforts to create a coherent and functional product, instead of designing them separately bazaar style then jerry rigging them together like some underfunded start-up making products for Kickstarter.