"pure fantasy thinking,"
"Today's humanoid robots will not learn how to be dexterous despite the hundreds of millions, or perhaps many billions of dollars, being donated by VCs and major tech companies to pay for their training,"
"we do not have such a tradition for touch data,"
"If the big tech companies and the VCs throwing their money at large scale humanoid training spent only 20% as much but gave it all to university researchers I tend to think they would get closer to their goals more quickly,"
Investors are pouring billions into humanoid robots while current humanoids remain coordination-challenged and lack real tactile dexterity. Human touch sensing is highly complex, with roughly 17,000 low-threshold mechanoreceptors concentrated at fingertips and about 15 neuronal families encoding pressure and vibration, creating a rich data modality absent from large-scale AI training. Training humanoids on videos of human tasks risks failing to produce true dexterity. Allocating a larger share of funding to university research could accelerate progress toward functional dexterity. Tesla and Figure continue rapid commercialization and fundraising despite unresolved tactile and coordination limitations.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]