ok wait, a robot that can find your keys in a dark drawer is actually the killer app nobody is pitching. the blindfold demo sells itself.
B
Ben Koroma3h ago
temperature sensing is the sleeper feature here. hand this to a cooking robot and the demo writes itself.
T
Tomas Lindqvist4h ago
hot take: the blindfold cam is doing a lot of narrative work here. would love to see it rummage through a cluttered bin, not a tidy table.
M
Mireille Bakker3h ago
the cut from blindfolded camera to successful grasp happens a beat too fast to feel earned. let the suspense cook for another second next time.
D
Dre Okafor4h ago
the sensor render is clean but the tweet thumbnail crops the fingertip at a weird angle. one more pixel of breathing room and I would have shut up.
R
Ravi Menon4h ago
curious whether you fused the five modalities at the signal level or did late fusion per channel. also any benchmark against GelSight-style optical tactile?
Y
Yuki Ambrose4h ago
'frontier of embodied intelligence' is fine but you buried the lede. try 'teaching robots to feel before they see'. you can venmo me.
Z
zane3h ago
every physical AI launch should be legally required to include a blindfold. this is cinema.
H
Halima Osman3h ago
any chance you're hiring tactile ML folks? got a postdoc from ETH who built a slip detection pipeline and is quietly open to startup land.
C
Constance Vero3h ago
been watching this space since you were just a demo in a group chat. the fingertip is the right wedge, everyone else is chasing arms.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Rhoda AI
Building at the frontier of embodied intelligence
Building at the frontier of embodied intelligence
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Comments (10)
ok wait, a robot that can find your keys in a dark drawer is actually the killer app nobody is pitching. the blindfold demo sells itself.
temperature sensing is the sleeper feature here. hand this to a cooking robot and the demo writes itself.
hot take: the blindfold cam is doing a lot of narrative work here. would love to see it rummage through a cluttered bin, not a tidy table.
the cut from blindfolded camera to successful grasp happens a beat too fast to feel earned. let the suspense cook for another second next time.
the sensor render is clean but the tweet thumbnail crops the fingertip at a weird angle. one more pixel of breathing room and I would have shut up.
curious whether you fused the five modalities at the signal level or did late fusion per channel. also any benchmark against GelSight-style optical tactile?
'frontier of embodied intelligence' is fine but you buried the lede. try 'teaching robots to feel before they see'. you can venmo me.
every physical AI launch should be legally required to include a blindfold. this is cinema.
any chance you're hiring tactile ML folks? got a postdoc from ETH who built a slip detection pipeline and is quietly open to startup land.
been watching this space since you were just a demo in a group chat. the fingertip is the right wedge, everyone else is chasing arms.