|
|
|
|
|
|
by JumpCrisscross
289 days ago
|
|
|
> there are some cool labs building foundational robotics models, but they're maybe ~5 years behind LMs today Wouldn't the Bitter Lesson be to invest in those models over trying to be clever about ekeing out a little more oomph from today's language models (and langue-based data)? |
|
|