Boston Dynamics shop: Spot Explorer

it's called "Inspection" and costs $29,750.00 :)

Explain

Counterpoint: No.

OP's humor might actually be too subtle for Reddit these days. They're busy recycling the same Black Mirror references.

This comment is specifically off-topic, per the rules. I don't, however, see jokes in comments mentioned.

> Please don't post comments saying that HN is turning into Reddit. It's a semi-noob illusion, as old as the hills.

https://news.ycombinator.com/newsguidelines.html

Threads like these make me wish for dang’s blessed “We’ve detached this thread...” comment.

Well it depends on what rmason plans to do with it.

I know what your saying, but I think the more of these are around the more freaked out I'm going to be.

Here's the video you're referencing. And yes, it is chilling in its utter believability.

https://www.youtube.com/watch?v=TlO2gcs1YvM

That's very well done, and not unimaginable either.

Thank you for sharing, that was fantastic.

There are a bunch of these competitions around the world:

https://en.wikipedia.org/wiki/Robot_combat

(It looks like RoboMaster is bigger in terms of the number of participants and spectators.)

There are a number of these types of collegiate competitions in China. A student was telling me about one he was working on where you had to optimize missile trajectories to avoid radar detection.

My Roomba has cliff detection, but it's probably easier to do in a rolling robot that's right next to the ground.

Yes, that surprised me, too. Cliff detection isn't that hard. They seem to have cameras in the nose area that are positioned to see whether there's a walkable surface ahead. The basic vision system is cameras with IR projection of a pattern, like the original Kinect. How can they do foot placement on stairs while unable to detect cliffs?

There's an optional LIDAR vaguely mentioned in the instructions. No specs on field of view or range, though.

"Spot’s joints can pinch fingers and other body parts and entangle loose clothing, long hair, and jewelry. Only handle Spot when the motors are locked out or robot power is off. Fingers may break or get amputated if caught in joints while Spot’s motors are active." Were pinch points not considered during design?

Jogging/trotting is only available in "demo mode", and is listed as "less stable".

Boston Dynamics can do better than this. I was expecting full Big Dog capability in a smaller package. Maybe this is just the minimum viable product for a market test. If someone finds a use case for this robot, the next model should have those problems fixed.

However, a very small child can ride it. :)

They did that with some Big Dogs, there's a video.

I guess more importantly, when an AT-AT [0] sized one?

[0] https://www.starwars.com/databank/at-at-walker/

Like BigDog? Not quite horse capacity, but still rideable in terms of load, although I don't know if it would screw up the balance

Seems like it'd be more effective to integrate a single beefier spot into the palanquin itself: https://steven-universe.fandom.com/wiki/Diamond_Palanquins

Wheeled robots already dominate indoor environments. They cost a tiny fraction of what this costs, have better endurance, and are safer.

> Flying drones have major endurance issues

What's the point of lasting 3 times longer if you are moving 5-10 times slower? Even if you double the endurance of the Spot, you are still covering the same amount of ground that a much more affordable drone could cover.

> observing dangerous environments like mines, places with toxic waste, and oil rigs.

This could be a good use. Mines in particular where drones would be impractical. I'm not so sure about oil rigs, while Spot would be great on the stairs/ terrain, it has trouble with seeing the edge of cliffs and might well march right off into the ocean.

Probably the biggest advantage would be the ability to take samples or hold some kind of special equipment.

As I understand/perceive it, current price point is aimed at R&D of use cases for when that price inevitably drops.

You got your top rated comment up-thread, now you're just double dipping!

Spot weighs 32.5 kg (71.7 lbs) according to https://shop.bostondynamics.com/DefaultStore/spot

So the A1 is 36% of Spot's weight.

Probably also features the plasticky hollow frail feeling of most chinese knockoffs. Still, if anybody wants to buy 7 of these and pit them against 1 Spot, I'd watch that :)

Even that is not a recurring cost if you can get the dog to "remember" the path taken with GPS coordinates.

At 30lb payload, it's not carrying any welder or compressor I've ever worked with. More like fetching a socket from the toolbox that's not quite in reach, but to accomplish those sorts of small and somewhat random tasks is going to require quite a bit more NLP than we've been able to do so far.

Easier said than parsed!

(As all Asimov's Robot stories pointed out)

Asimov's laws are only good for a lifetime writing career in exploring their exceptions. Is there any Asimov story where the laws were applied strictly and successfully? I don't remember one.

They don't brick them, they impound them.

For lesser violations they can make it temporarily play dead.

Selected Q/A from the 'SPOT PRIVACY NOTICE' [1].

Of interest is that access to the Performance Log is required at "license renewal time [...] before a new key will be issued."

I've not seen other info about their licensing terms but am curious to read those if somebody has a link to those details.

According to the Spot Privacy Notice, Performance Log includes: - Your robot’s serial number. - Information about usage such as how long the robot has been operated, how far it walked, how many times it fell, how often a particular API was used. - When key events like powering on or fault detection occurred. - Safety-related data such as collisions or falls. - Usage of the controller application. - High-level information about maps you collect, such as how many waypoints, how they’re connected, how well Spot understands where it is. - Details about communications loss events.

Q: Who will have access to my data? A: Access to your data will be limited to Boston Dynamics customer support and individuals related to improving the product.

Q: Under what circumstances is Spot data transmitted to Boston Dynamics? A: Service Logs: Only when explicitly sent by you as part of a service request. Performance Logs: Regularly and automatically, whenever the tablet is connected to the Internet. [emphasis added]

Q: Does Boston Dynamics retain my data? A: Service Logs: Deleted when a service ticket is resolved. Performance Logs: Yes, this data is retained as long as it is necessary for performance improvements to Spot robots.

Q: What if I don’t use the tablet or Android app with Spot? A: If you have a Spot Explorer we’ll ask you for your Performance Log at license renewal time. At that point you will have to use Spot’s Android App to transfer the log before a new key will be issued. [emphasis added]

Q: Does Spot require Internet access? A: No. In order to facilitate transmission of diagnostic data we do require you to periodically connect your tablet to the robot, which automatically transfers the data from the robot, and then connect that tablet to the Internet.

[1] https://www.bostondynamics.com/spot/privacy-notice

I would hazard a guess of proprietary form factor and handling needs since these aren't mounted in a car chassis, combined with low production volumes initially.

You might get away with 3 (depending on load), but then you would need an additional charger ($1360 each). But there was a reason I use the ~ symbol =). Personally if I were doing this I would get more chargers, but it is really up to the use case. Either way, assuming 5,000 hours/year, you are looking around $20-30/hour of effective cost (depending on what you count for “everything else)

I'm sure there is already a Chinese made clone available that used stolen IP.

Well they have $20,000+ addons for better Lidar range, on device compute etc. Doesn't seem all that cheap to do interesting things..

Not sure this take feels right but it doesn't really feel sufficient to say "this time is different" either. It's pretty easy to come up with dozens of military based use cases but really difficult to find commercial ones. Possible use cases for say delivery still rely on human drivers (in delivery vans) or tele-operators for example so it's not clear there's a business case for using them.

This seems like the mother of all selection biases to me. Of course you remember the Apple ][ because it was one of the first products in what became a wildly successful market. How many flying car prototypes do you remember? There have been several, and a whole new batch of electric versions are being designed right now, but so far people have been quite rightly skeptical of that product for the last 70+ years.

I guess. It still makes me wonder about the margins for other accessories (let alone Spot itself).

Max payload is only 31lbs unfortunately

Entertainment, aka "we don't know we just think it's cool"

The only way this picture could get any better was if the house in the background was exploding.

Except someone would try to steal/destroy it and it couldn’t do anything, so not really. Maybe in a decade or two.

Did Royal Caribbean ever get the robotic bartenders to actually work? When I got to see one in action several years ago it spent far more time either outright down or in a half broken state where it messed up orders than it did actually working. Not to mention it being much slower than a human bartender when it was operational.

Well, some technology can't be replicated by living creatures... I could hire 10,000 book-keepers and they still wouldn't be as fast or as accurate at handling my company's finances than an old pirated copy of Quickbooks.

On the other hand, would I spend $7,000 on a state of the art sex doll? I'd probably just head to the bar, spend $50 or so on drinks, and take home a female human.

But, I'm cheap...

Some will spend the $75k to have an automated quadruped robot fetch them their sex doll. As for me, I'll just bring my little chihuahua to the bar, use him to flirt with some girl, and head home with the dog, the girl, and some beer all for much less than $75k.

It's harder to get actors to respond realistically to an animated robot than a physical one. Depending on what you want, I could well see this being cheaper and easier. At the limit, consider if you only need stuff it can do out of box, and BD will lend you one for promotional purposes or you can resell it with minimal depreciation.

Then why would Disney pile money into an acrobatic robot? I think there must be a use case there somewhere

Judging by the other comments ("Finally! Something to retrieve a socket wrench from all the way across the room!"), my view that this robot is utterly useless is indeed an unpopular opinion. This is HN and people love the "oh look this is so tech" kind of story. The whole startup scene is about it. I'm not seeing a ton of other people who said the same thing prior to my comment, but to placate you I edited my comment to remove the "unpopular" part. I would ask that you reply to what I say, rather than the specific wording. If I may point it out, your comment critiqued my choice of words while totally ignoring the point I stated.

The argument is: these things are useless in the real world.

This argument is falsifiable: come up with a use case for this robot that is real-world reasonable given the price point and other practical considerations.

If BD is charging $75k for a product that has been in development im one form or another for 10+ years, it seems reasonable that the burden is on BD to prove it's a useful product.

Judging by the history of the product, it seems clear that it's a novelty item. The DoD turned its back on the very loud BigDog, and Google handed BD off to Softbank when they weren't able to answer the question, "so, how do we make money from selling this robot to consumers?"

You can buy three Ford pickup trucks, brand spanking new, for the price of one of these robots.

Boston Dynamics isn't exactly a company that makes products. They are a bunch of researchers who turn money into technology. That technology may involve impressive robotic motion, but not every impressive thing is useful to consumers.

There's a very strong sense among many robotics weenies at MIT that, if you make cool stuff, the DoD or some major corporation will eventually reward you with money that allows you to make more cool stuff. That sense is embodied by Boston Dynamics, in my mind. It doesn't translate to a useful product, whether intended for the consumer, commercial applications, or heavy industry.

And the dangerous bit is that this engineering philosophy of "find someone who will pay us to keep making cool useless stuff" only serves to create tools for the over-funded, which in our society means militaries, police, and other anti-human actors. So the most likely use of this product, if there is any use at all, is a destructive and anti-people application. But I don't think it's very useful even for those purposes, because of price, fragility, need to recharge, light payload, etc.

Be honest. Can you think of any specific application where this product is better than something else (including human labor) that can be bought for $75k?

I know it's cgi, their channel is pretty cool, check out Bohemian Rhapsody: https://youtu.be/PJO7bcRVPvI

Had no idea about Google, seems like the perfect moonshot, wonder what they were expecting.