2 ms·

I'd guess that the giant agricultural firms have already built quite accurate models using decades of data and expertise, and the addition of fancy new AI techn

I'd guess that the giant agricultural firms have already built quite accurate models using decades of data and expertise, and the addition of fancy new AI techniques didn't offer large improvements.

TaylorAlexander 2y ago
As someone who worked at Google X robotics until leaving to start an open source farming robotics project very similar to Mineral (see my profile) I can say that there’s so, so much opportunity for new ideas in this field, but Google X is not very good at launching actual companies. So much of what’s done at Google X is canceled (I worked on what became Everyday Robots, now canceled), that you can’t take “project at X was canceled” as a sign of anything at all really. There are very weird forces going on at Google X. I don’t think I’m equipped to provide a thorough analysis, but I think there’s just too much easy money and the only people they really have to please are the guys at the top, who are deca billionaires who don’t even have to spend their own money to fund this stuff. It’s all a very fun and interesting R&D adventure, and those guys really have accomplished a lot in this life so I wouldn’t write off their abilities by any means, but the arrangement is very strange by traditional R&D standards. And it really seemed to me like the people in charge of our project were more focused on getting massive quarterly milestone bonuses than they were actually delivering something that would work. Honestly I really can’t say exactly what is going wrong, except that Google X is a very strange place.
- pstuart 2y ago
  Acorn looks super cool. Fight the good fight!
  - TaylorAlexander 2y ago
    Thank you so much! I made some good progress tonight. Finished bringing up the system with all of our new electronics, and in the next six months we’re going to start building a small fleet of them!
- killerstorm 2y ago
  I guess Google is not interested in small things. E.g. $1M revenue might sustain a startup in other parts of the world, but for Google it's a rounding error.

I'd guess that the giant agricultural firms have already built quite accurate models using decades of data and expertise, and the addition of fancy new AI techniques didn't offer large improvements.

TaylorAlexander 2y ago

As someone who worked at Google X robotics until leaving to start an open source farming robotics project very similar to Mineral (see my profile) I can say that there’s so, so much opportunity for new ideas in this field, but Google X is not very good at launching actual companies. So much of what’s done at Google X is canceled (I worked on what became Everyday Robots, now canceled), that you can’t take “project at X was canceled” as a sign of anything at all really. There are very weird forces going on at Google X. I don’t think I’m equipped to provide a thorough analysis, but I think there’s just too much easy money and the only people they really have to please are the guys at the top, who are deca billionaires who don’t even have to spend their own money to fund this stuff. It’s all a very fun and interesting R&D adventure, and those guys really have accomplished a lot in this life so I wouldn’t write off their abilities by any means, but the arrangement is very strange by traditional R&D standards. And it really seemed to me like the people in charge of our project were more focused on getting massive quarterly milestone bonuses than they were actually delivering something that would work. Honestly I really can’t say exactly what is going wrong, except that Google X is a very strange place.

pstuart 2y ago
Acorn looks super cool. Fight the good fight!
- TaylorAlexander 2y ago
  Thank you so much! I made some good progress tonight. Finished bringing up the system with all of our new electronics, and in the next six months we’re going to start building a small fleet of them!
killerstorm 2y ago
I guess Google is not interested in small things. E.g. $1M revenue might sustain a startup in other parts of the world, but for Google it's a rounding error.