Lighting differences can make a huge difference in robotics. Today, I found a quirk in my model exemplifying this.
> I collected 10h of training data.
> 3h in, I notice that the left arm following the right arm for the final movement could be good for the final insertion subtask.
> I change behaviour.
> But the first 3h didn't include much nighttime data (=dark room).
🎯Result
The same policy reliably exhibits different behaviour on exactly this subtask depending on the lighting.
So many microlearnings to be made in robotics.
Webb’s images from Tj are special to me bc they show the world I was born into and grew up in. I still cross the same border to get to work and still see the curio guys do their rounds. Doesn’t hurt that Webb has more juice than entire generations of street photographers combined
Software guys who never work on embedded stuff love saying shit like “oh I’m just updating the firmware on my freaking espresso machine. Totally normal thing to do!” Yeah man everything is computers now and they run software you should know this already you work in tech