Hacker News Reader

Comment by xpl | Hacker News Reader

> They will always be expensive, bulky, and creepy.

People said similar things about those early NNs generating images, code, music.

Surely we will solve the fine motor skills soon. It just takes one incremental step at a time, there are no fundamental roadblocks on the way.

> No one will want one in their homes.

I would kill for a robot that keeps my house clean, makes my favorite food, fetches me things, etc.

While I could pay a human to do that, it just seems more convenient if that wasn't a real person. Because with a real person I can't e.g. walk naked around them and behave generally weird, because I need to respect their feelings, which I don't have to do with a robot.

View on HN

> > They will always be expensive, bulky, and creepy.

> People said similar things about those early NNs generating images, code, music.

Yes and… they still are expensive, bulky, and creepy. The expense and bulk is externalised to the stock market (but maybe improving), the creepy is endemic and not getting any better.

https://www.midjourney.com/showcase is funded by customers, not investors. Anything in there that's creepy, but not obviously intentionally creepy, is more a reflection on the viewer than the image :P

Not sure walking naked in front of an internet-connected camera in humanoid form is much better...

I would pay extra for a robot that runs on a local hardware — there would be a market for that, for sure. People need privacy.

An expensive/bulky GPU doesn't have to be built in the robot body itself — it could be a box mounted somewhere in the house with the wireless connection to the robot body.

I am sure at some point there will be hardware that would protect model vendors from extracting/copying the model weights, like on-chip encryption when doing inference or something.

Or maybe there will be GPUs with homomorphic encryption¹, so everything will be end-to-end encrypted, and it will be safe to compute in the cloud without compromising user privacy.

[1] https://en.wikipedia.org/wiki/Homomorphic_encryption

The NVIDIA keynote today mentioned on-chip encryption for model data actually. Not sure how entirely that covers your proposed case, but it’s already something they’re working on at least.

Reply on news.ycombinator.com