And also, to those that know, is what Google is doing significantly market leading? It seems that a lot of other robot demos are about jumping around over motor control.
"Attention is All You Need" is a research paper published by google in 2017 that led pretty directly to the AI/LLM boom we have today.
Sometimes the pipeline between SOTA research is longer than you want, but it doesn't mean that publishing research or doing research should be discouraged.
I rather appreciate that google is willing to show their work, and throw money at things that might make the world a better place without a clear path to profitability.
For all their faults, their willingness to support pure research isn't one of them.
Google does some stellar research as another commenter has pointed out.
Sure, anything that the shareholders or executives can get their grubby, misinformed endless growth paws into withers and dies, but that's the modern life we all vote for.
To those in the field, how long until it can navigate a zipper?
And also, to those that know, is what Google is doing significantly market leading? It seems that a lot of other robot demos are about jumping around over motor control.
Doesn't look like it, benchmarks in article have the not-ER model slightly below GPT-5, which is puzzling.
seems none of the existing models focus on aerial robots
What would you like aerial robot to do? Navigation, tracking, delivery sort of done by algorithms + low level recognition
all of that :)
If you want the latest ai for drones then ukraine has a lot of training data.
”guys we did research!”
Is it a product you can buy or a thing you can use? Of course not, this is Google.
"Attention is All You Need" is a research paper published by google in 2017 that led pretty directly to the AI/LLM boom we have today.
Sometimes the pipeline between SOTA research is longer than you want, but it doesn't mean that publishing research or doing research should be discouraged.
I rather appreciate that google is willing to show their work, and throw money at things that might make the world a better place without a clear path to profitability.
For all their faults, their willingness to support pure research isn't one of them.
Google does some stellar research as another commenter has pointed out.
Sure, anything that the shareholders or executives can get their grubby, misinformed endless growth paws into withers and dies, but that's the modern life we all vote for.
This model is available for me to prompt in Google’s AI Studio as of 1-2 days ago.
There’s literally a link to a thing you can use. Google 2025 has some great PMs and dev rel.