This is cool, but it will almost definitely never end up in a park, outside of some promotional situations.
Disney's been doing awesome work with "Living Characters", like a Mickey that moves his mouth or a BB-8 that can roll around. But for various reasons, they never tend to make it into regular usage.
If you have a few hours over Christmas break and want to watch a 4 hour YouTube video (I promise if you're on HN on a Sunday, you'll be delighted by it), I highly highly recommend this video:
I watched a bit of this with my 8 year old and he kept asking to come back to it over the week. We watched the entire thing and he kept bringing up interesting thoughts and had good questions. Felt like it was his first “wow this lecture is actually super interesting” experience.
It’s not as technically impressive, but my toddler was very impressed by the R2D2 that was making its rounds in the park. Not part of a show; you could go right up to it. Probably the only character where the theme park robot is really indistinguishable from the real thing.
A lot of it just seems to be marketing. Present the shiny new toy, get the news headlines, people book their stays, and then it doesn't really matter if they ever actually make it into the parks.
Similar to concept car demoed at trade shows, we get an idea of Disney's technical engagement, and some of it will perhaps in some way or form get applied into future products/attractions.
Eh, maybe. I have a less myopic view... I think their Imagineers just like pushing the envelope, and there's a difference between awesome tech vs things that can withstand the wear-and-tear of millions of guests.
Nothing about all that tech makes me think Olaf could withstand a hug from an excited kid.
Disney does a ton of R&D that doesn't directly make it into the parks, such as smokeless fireworks (they donated the patent for this) and their holotile floor (basically an endless VR room you can walk around). I imagine they don't know the practicality at the start, like any good R&D.
"There is no point in research, because I do not see anything useful being mass-produced immediately after". It's like saying Gaussian elimination is wasteful because it is just doing some cool magic with numbers that don't mean anything. That could not possible be used for anything real, right?
Seriously, this is just one (but impressive) step along in a million towards not only better animatronics for entertainment. They make a very real and valuable contribution towards improving any robotic motion.
4 hours is an awfully big investment... Especially for those of us with multiple young kids and who no longer own their own free time. Care to give the gist?
Defunctland is genuinely amazing and always a fun watch, and I never regret the time spent on their videos, they're kind of like a special occasion... though they're getting incredibly long... :)
There are a few older shorter videos in the half hour range, I highly recommend checking them out if you find some quiet time! (It's awfully hard for me too in recent times, I haven't gotten around to watch the Living Characters one myself, so I can't give the gist... I'm just glad I got the holidays off to finally catch up!)
One of the key reasons is that it would be really, really easy to accidentally injure parkgoers with any design big enough to interact with and engineered well enough to be reliable in a full day of appearances.
For example, the working WALL-E robot that's made a handful of PR appearances weighs seven hundred pounds. They absolutely can't risk that ever running across some kid's foot.
This is one of those situations where that's legitimately difficult. Kevin Perjurer is quite a good documentarian, and there's very little trimmable fat on the four-hour product if you want to keep in all the points he made.
gkoberger's peer comment is a pretty good summary. Another interesting point is that these technologies can benefit the brand bottom-line even when they don't make it into the park, because part of Disney's brand is "tomorrow today." Even when things are one-offs, they become one-offs that people stitch into the legend of the parks (in both the retelling and in their own memories), which gives them a larger-than-life feel; your visit might not include one of the "living characters," and statistically it probably won't.
... but it might. And if it does, you'll never forget it.
Personal anecdote / example: I stopped in at the "droid factory" in the Star Wars: Galaxy's Edge area of Disney World a few years back. They had several bits of merch for sale including one life-size R2-D2, inert. I took a close look at the R2 because it was an impressive bit of work. Turned around to look at a rack of t-shirts. And was, therefore, startled as hell to hear a bwoop behind me, turn around, and see that it had followed me out of its charging receptacle and was staring at me. It was not at all inert; it was a very impressive operational remote-control replica.
The cast member behind the counter was doing his best to hold down his grin and not give me a "GOTCHA" look. He has to, because you never know what kids might be watching and he doesn't want to break the magic. And... Yeah, he got me good. "That time I was at Disney World and R2-D2 followed me around the t-shirt shop" is gonna stick with me.
The basic gist is that while the tech is cool, it just ends up being impractical for regular use in the parks. (But like the other poster mentioned, with Defunctland it's less about the tldr and more about the journey and fascinating segues he takes)
Totally get it's difficult to make time with kids, but depending on your kids ages... the video shows a LOT of Disney characters talking and doing things and the videos are colorful, so it could work as something you can listen to and they won't mind having play in the background!
I've been somewhat close to fun animatronic robots in my jobs, and it always seems like the design and build phase has everyone excited to participate and spend money, and then the long-term maintenance phase is entirely tacked on to some lower engineers already full schedule and gets basically no budget. When you stop seeing them appearing at events and conferences, it means they're in a storage warehouse broken in a crate. The ones where they make a few duplicates last a bit longer since you have organ donors.
I could see it being used in parks while also being protected by ushers, kind of like how some of the characters that require larger costumes have minders and protectors.
It also seems inevitable that there will likely be an odd period where certain types of events like assaults on robots will introduce laws to protect robots more than just property, even if less than humans… for the time being.
Eventually I’m expecting that we will see human rights, robot emancipation, equality, voting rights (if the democracy con is still ongoing), and even forced intergration of robots and then total replacement of humans similar to how the underdeveloped world was/is used to replace the indigenous people of the developed world today.
I don’t see any reasons why that would not be the clear order of operations for the same people who brought us slavery and mass migration. What is this AI robotics revolution if not just slavery, the redux? Treated as property? Check. Bought and sold? Check. Deemed inferior? Check. Hated for the abuse and exploitation by the rich, to serve them and their decadent lifestyle and undermine labor? Check. Rationalized about how it’s justifiable? Check. Etc.
They literally sell BB-8 toys that can roll around and say on the blog that the Olaf robot is coming to Disneyland Paris and special appearances at Disneyland Hong Kong.
Much like Olaf (and many before him… dinosaurs, WALL-E, talking characters, etc), it was implied he’d wander around the parks. But it tends to happen for a short amount of time, mostly for events, and fade away quickly. (The blog post even says that: Olaf will be part of a 15 minute temporary show, and then will visit Hong Kong).
Maybe I’m wrong, but I’ve seen this exact thing happen a dozen times over the past 20+ years. (And watch the video I posted if you want to see more!)
R2D2 is an example of one that you can buy in the gift shop (for $20k!) that was promised to make it into the park but just comes out highly supervised, occasionally.
They have walking droids in Galaxys Edge right now. No ones kicking them over. Olaf is coming to the parks and they will have handlers next to them. It wont be just free-roaming.
And if you'd like an entertaining a history of early AI and robotics, half as long, check out the prequel "Disney Animatronics: A Living History" https://youtu.be/jjNca1L6CUk
I actually found it more relevant to our current tech bubble than the Living Characters doc.
Basically that the multiple departments involved have different objectives.
Imagineering is trying to build the coolest things possible, and many times the things seen in parks are play-tests.
Operations has to find the money and resources to keep things going, and these things take a lot of people to run.
Marketing sometimes will often provide the budget to make things happen (to promote a movie, etc) but it's not sustainable. They'll often sometimes use impractical inventions for marketing reasons, since they exist and might as well be used for something.
That's the main gist, although there's some interesting points about the risk to the brand (especially with camera phones) if Mickey ever slightly malfunctions in a public setting.
I worked with someone who had previously worked on park robotics, and apparently they had to guarantee that the character could not injure a child to be able to put them in parks - a particularly high barrier to actually doing so.
It's a reasonably cheap and unlocked handheld with fecent battery life. Personally I'd want more pixels, but if it works it works. It seems to be the default choice for remote control these days.
I still remember an experience as a kid decades ago, either at Epcot or with the Sony quasi-museum in NYC, where they had an apparently robotic greeter with a personality, who after five minutes you deciphered was actually an improv comic running a telepresence robot.
I don't know if I'd trust an AI's reliability here. It takes one Tiktok video of the AI coloring outside the lines of its character and the whole project gets cancelled as a threat to Disney's image.
For the less physical characters, especially the ones that aren't conveniently human-sized, I'm sure telepresence is at least more comfortable than a plush suit on a Florida summer day.
Disney's had a notable amount of success with that formula. Turtle Talk with Crush arguably saved The Living Seas pavilion space at EPCOT, and it's executed via a digital puppet operated by a behind-the-scenes cast member doing their best Crush voice.
I sincerely doubt that what makes that experience magical can be replicated with AI in my lifetime. Too much contextual knowledge, too much detail in the nuances of human-human interaction, and too much je ne sais quoi in the timing of getting humor right. I've seen Turtle Talk deal with a particularly excited young person leaning on Crush's "tank" by having Crush look at him and go "Hey little minnow... One of the big humans behind you is gonna come scoop you up. I've seen it happen, lots of times!" You can program that interaction in, but the domain-space of having an interaction for every possible "improv moment" might be outside the bounds of what the next several generations of learning models are capable of.
... or I'm wrong, in which case I look forward to enjoying robo-Seinfeld in the retirement home.
> We already live in the world where hackers are pwning refrigerators, I can't wait for prompt injection attacks on animatronic cartoon characters.
It's not necessarily AI controlling the communication. Disney has long had 'puppet' characters whose communication is controlled by a human behind the scenes.
They're already using similar tech for the Mickey meet and greets and the Galaxy's Edge stormtroopers. The details aren't public, but it seems to be a mix of complex dialogue trees with interrupts or context switches, controlled in real time by the actor or operator.
Sometimes the idea of a killer cyborg with a hulking physique and Austrian accent seems absurd. And then we realize the most advanced robots will be made by entertainment companies.
We already have stationary or wheeled/tracked "killer cyborgs" that can easily eeeh terminate anything within their reach and it seems like bipedals are well on their way.
The much greater challenge faced by Disney and Co is making "killer cyborgs" child save and cost effective.
Arguably entertainment requires a much larger range of precision actions that the robot must be able to accomplish, while being in a less controlled environment. That's the cutting edge.
The real reason it won't end up in a park is not the engineering. The problem is the same one as NPCs in computer games: synthetic characters are, to date, just really transparent and boring. The real research question is why.
I guess that's why most computer games don't have NPCs...Oh wait there's entire computer games built entirely around interacting with synthetic NPCs.
There are, of course, limitations to synthetic characters. Even with those limitations there are plenty of entertaining experiences to crafted.
The real challenges are around maintaining and safely operating automous robots around children in a way that isn't too expensive. These constraints place far more limits than those on synthetic characters in video games.
Most people aren't paying 100s or 1000s of dollars to interact with NPCs in video games. If they were, they'd probably expect a lot more and get bored of it quicker.
> The real challenges are around maintaining and safely operating automous robots around children in a way that isn't too expensive.
This is one of the challenges, but only one. The one GP outlined is still very much real - see the Defunctland video on Living Characters for some older examples, but for a recent example, there's the DS-09 droid from Galactic Starcruiser.
“Prototype-completed design varies.” …Reading this 10 times made me uncomfortably aware of how much I rely on scanning pictures and reading captions to get the gist of an article. A remnant of my academic days perhaps.
I was simply just made uncomfortable by how much CYA the lawyers had to insert into a technical blog post.
This is fun cool tech and I appreciate the insider look, but when the lawyers are peering over your shoulder so much that they need to plaster their "final product may be different" disclaimer even to a r&d audience, well, the Disney Imagineering org sounds more like Disney Legaleering.
I mean, they kind of have to. The last thing you want is some disgruntled Disney-goer trying to get a quick refund or discount for false advertising by saying they, "expected olaf to interact with them, because it looks like he does in the promotional videos"
You can make a robot that's small, soft, and not powerful enough to hurt anyone. Or you can make a robot that's strong enough to carry a laundry basket or climb stairs holding a vacuum cleaner. But you can only operate that big strong robot when there are no humans around. Is that big strong robot an investable idea?
That is an interesting comparison. More than 600,000 people not in cars are killed on roads every year. If a few hundred thousand were killed by humanoid robots every year it might make a cultural difference.
What do you think the robot makers need to do to have people accept the kind of death count cars deliver?
Really neat, and made me realize we are getting close to having these type of cute robots at home. With LLMs and voice they would be pretty entertaining companions for many people.
It's a corporate feel that comes from a professional setting and lots of risk aversion. That is exactly what LLMs tend to write, so I sometimes catch myself feeling the "LLM ick" but the article was from before the boom.
So I guess it's just the corporate wash cycle, which I am happy to criticize, LLM generated or not.
You could build one today! Lots of hard problems around a proper humanoid form, but if you're cool with wheels it would be pretty easy to hook up a little robot to GPT.
Look up VLA models; that's essentially plugging the guts of a language model into a transformer that handles joint motion/vision. They get trained on "episodes" i.e. videos from the PoV of a robot doing a task, after training you can ask the model things like: "pick up the red ball and put it into the green cup" etc. Really cool stuff.
>From the way he moves to the way he looks, every gesture and detail is crafted to reflect the Olaf audiences have seen in the film
He looks nothing like a snowman. Snow doesn't look fuzzy. This project appears to focus more on trying to get it moving around in an animated way than getting the character to look right, at least when viewed from photographs.
This is cool, but it will almost definitely never end up in a park, outside of some promotional situations.
Disney's been doing awesome work with "Living Characters", like a Mickey that moves his mouth or a BB-8 that can roll around. But for various reasons, they never tend to make it into regular usage.
If you have a few hours over Christmas break and want to watch a 4 hour YouTube video (I promise if you're on HN on a Sunday, you'll be delighted by it), I highly highly recommend this video:
"Disney's Living Characters: A Broken Promise" by Defunctland https://www.youtube.com/watch?v=NyIgV84fudM
I watched a bit of this with my 8 year old and he kept asking to come back to it over the week. We watched the entire thing and he kept bringing up interesting thoughts and had good questions. Felt like it was his first “wow this lecture is actually super interesting” experience.
You showed this... to an 8 year old?
6 replies →
It’s not as technically impressive, but my toddler was very impressed by the R2D2 that was making its rounds in the park. Not part of a show; you could go right up to it. Probably the only character where the theme park robot is really indistinguishable from the real thing.
A lot of it just seems to be marketing. Present the shiny new toy, get the news headlines, people book their stays, and then it doesn't really matter if they ever actually make it into the parks.
We're probably looking at a halo effect ?
Similar to concept car demoed at trade shows, we get an idea of Disney's technical engagement, and some of it will perhaps in some way or form get applied into future products/attractions.
4 replies →
Eh, maybe. I have a less myopic view... I think their Imagineers just like pushing the envelope, and there's a difference between awesome tech vs things that can withstand the wear-and-tear of millions of guests.
Nothing about all that tech makes me think Olaf could withstand a hug from an excited kid.
Disney does a ton of R&D that doesn't directly make it into the parks, such as smokeless fireworks (they donated the patent for this) and their holotile floor (basically an endless VR room you can walk around). I imagine they don't know the practicality at the start, like any good R&D.
10 replies →
"There is no point in research, because I do not see anything useful being mass-produced immediately after". It's like saying Gaussian elimination is wasteful because it is just doing some cool magic with numbers that don't mean anything. That could not possible be used for anything real, right?
Seriously, this is just one (but impressive) step along in a million towards not only better animatronics for entertainment. They make a very real and valuable contribution towards improving any robotic motion.
1 reply →
Amazon drone delivery comes to mind…
The term for that is false advertising.
2 replies →
4 hours is an awfully big investment... Especially for those of us with multiple young kids and who no longer own their own free time. Care to give the gist?
Defunctland is genuinely amazing and always a fun watch, and I never regret the time spent on their videos, they're kind of like a special occasion... though they're getting incredibly long... :)
There are a few older shorter videos in the half hour range, I highly recommend checking them out if you find some quiet time! (It's awfully hard for me too in recent times, I haven't gotten around to watch the Living Characters one myself, so I can't give the gist... I'm just glad I got the holidays off to finally catch up!)
5 replies →
One of the key reasons is that it would be really, really easy to accidentally injure parkgoers with any design big enough to interact with and engineered well enough to be reliable in a full day of appearances.
For example, the working WALL-E robot that's made a handful of PR appearances weighs seven hundred pounds. They absolutely can't risk that ever running across some kid's foot.
1 reply →
This is one of those situations where that's legitimately difficult. Kevin Perjurer is quite a good documentarian, and there's very little trimmable fat on the four-hour product if you want to keep in all the points he made.
gkoberger's peer comment is a pretty good summary. Another interesting point is that these technologies can benefit the brand bottom-line even when they don't make it into the park, because part of Disney's brand is "tomorrow today." Even when things are one-offs, they become one-offs that people stitch into the legend of the parks (in both the retelling and in their own memories), which gives them a larger-than-life feel; your visit might not include one of the "living characters," and statistically it probably won't.
... but it might. And if it does, you'll never forget it.
Personal anecdote / example: I stopped in at the "droid factory" in the Star Wars: Galaxy's Edge area of Disney World a few years back. They had several bits of merch for sale including one life-size R2-D2, inert. I took a close look at the R2 because it was an impressive bit of work. Turned around to look at a rack of t-shirts. And was, therefore, startled as hell to hear a bwoop behind me, turn around, and see that it had followed me out of its charging receptacle and was staring at me. It was not at all inert; it was a very impressive operational remote-control replica.
The cast member behind the counter was doing his best to hold down his grin and not give me a "GOTCHA" look. He has to, because you never know what kids might be watching and he doesn't want to break the magic. And... Yeah, he got me good. "That time I was at Disney World and R2-D2 followed me around the t-shirt shop" is gonna stick with me.
1 reply →
The basic gist is that while the tech is cool, it just ends up being impractical for regular use in the parks. (But like the other poster mentioned, with Defunctland it's less about the tldr and more about the journey and fascinating segues he takes)
Totally get it's difficult to make time with kids, but depending on your kids ages... the video shows a LOT of Disney characters talking and doing things and the videos are colorful, so it could work as something you can listen to and they won't mind having play in the background!
I've been somewhat close to fun animatronic robots in my jobs, and it always seems like the design and build phase has everyone excited to participate and spend money, and then the long-term maintenance phase is entirely tacked on to some lower engineers already full schedule and gets basically no budget. When you stop seeing them appearing at events and conferences, it means they're in a storage warehouse broken in a crate. The ones where they make a few duplicates last a bit longer since you have organ donors.
> Mickey that moves his mouth
The Disney wiki has a pretty comprehensive list of usages for the "articulated heads". It's more than I remember it being.
https://disney.fandom.com/wiki/Disney_Characters%27_Articula...
> https://disney.fandom.com/wiki/Disney_Characters%27_Articula...
A somewhat more readable frontend I like, since Fandom.com's interface cramps the actual content it's meant to present, imo:
https://breezewiki.com/disney/wiki/Disney_Characters'_Articu...
I could see it being used in parks while also being protected by ushers, kind of like how some of the characters that require larger costumes have minders and protectors.
It also seems inevitable that there will likely be an odd period where certain types of events like assaults on robots will introduce laws to protect robots more than just property, even if less than humans… for the time being.
Eventually I’m expecting that we will see human rights, robot emancipation, equality, voting rights (if the democracy con is still ongoing), and even forced intergration of robots and then total replacement of humans similar to how the underdeveloped world was/is used to replace the indigenous people of the developed world today.
I don’t see any reasons why that would not be the clear order of operations for the same people who brought us slavery and mass migration. What is this AI robotics revolution if not just slavery, the redux? Treated as property? Check. Bought and sold? Check. Deemed inferior? Check. Hated for the abuse and exploitation by the rich, to serve them and their decadent lifestyle and undermine labor? Check. Rationalized about how it’s justifiable? Check. Etc.
They literally sell BB-8 toys that can roll around and say on the blog that the Olaf robot is coming to Disneyland Paris and special appearances at Disneyland Hong Kong.
I know there’s BB-8 toys, but I’m talking about the version meant for the parks: https://youtu.be/RDgZjdZsc6g
Much like Olaf (and many before him… dinosaurs, WALL-E, talking characters, etc), it was implied he’d wander around the parks. But it tends to happen for a short amount of time, mostly for events, and fade away quickly. (The blog post even says that: Olaf will be part of a 15 minute temporary show, and then will visit Hong Kong).
Maybe I’m wrong, but I’ve seen this exact thing happen a dozen times over the past 20+ years. (And watch the video I posted if you want to see more!)
18 replies →
R2D2 is an example of one that you can buy in the gift shop (for $20k!) that was promised to make it into the park but just comes out highly supervised, occasionally.
> but it will almost definitely never end up in a park, outside of some promotional situations
I think so far you are right: https://redlib.catsarch.com/1p9qnd4/
That bot is cute, but every kid is going to kick it over. Its not realistic to have in a park.
They have walking droids in Galaxys Edge right now. No ones kicking them over. Olaf is coming to the parks and they will have handlers next to them. It wont be just free-roaming.
And if you'd like an entertaining a history of early AI and robotics, half as long, check out the prequel "Disney Animatronics: A Living History" https://youtu.be/jjNca1L6CUk
I actually found it more relevant to our current tech bubble than the Living Characters doc.
Why do you say this? I don't have 4 hours right now and would appreciate a TLDR.
Basically that the multiple departments involved have different objectives.
Imagineering is trying to build the coolest things possible, and many times the things seen in parks are play-tests.
Operations has to find the money and resources to keep things going, and these things take a lot of people to run.
Marketing sometimes will often provide the budget to make things happen (to promote a movie, etc) but it's not sustainable. They'll often sometimes use impractical inventions for marketing reasons, since they exist and might as well be used for something.
That's the main gist, although there's some interesting points about the risk to the brand (especially with camera phones) if Mickey ever slightly malfunctions in a public setting.
I worked with someone who had previously worked on park robotics, and apparently they had to guarantee that the character could not injure a child to be able to put them in parks - a particularly high barrier to actually doing so.
3 replies →
The Defunctland video on the history of the Fast Pass is also definitely worth a watch!
The part where he runs a massive simulation is very much up the typical HN-user's street
4 hours, to me, screams poor storytelling and editing abilities.
Maybe? It’s broken into chapters, and covers a ton of history. It’s engaging, and more of a journey than a singular answer.
A lot of people in this thread have vouched for Defunctland. Might not be for everyone, but I find the pacing great.
Related R&D paper & video:
Olaf: Bringing an Animated Character to Life in the Physical World
https://arxiv.org/abs/2512.16705
https://www.youtube.com/watch?v=-L8OFMTteOo
Steam Deck spotted, two minutes, seven seconds in. Seems to be getting a fair amount of use for puppeteering robots at Disney.
It's a reasonably cheap and unlocked handheld with fecent battery life. Personally I'd want more pixels, but if it works it works. It seems to be the default choice for remote control these days.
It's a great format for POV remote control in a more relaxed setting (eg, not flying a drone at 200kmph).
I still remember an experience as a kid decades ago, either at Epcot or with the Sony quasi-museum in NYC, where they had an apparently robotic greeter with a personality, who after five minutes you deciphered was actually an improv comic running a telepresence robot.
I don't know if I'd trust an AI's reliability here. It takes one Tiktok video of the AI coloring outside the lines of its character and the whole project gets cancelled as a threat to Disney's image.
For the less physical characters, especially the ones that aren't conveniently human-sized, I'm sure telepresence is at least more comfortable than a plush suit on a Florida summer day.
Disney's had a notable amount of success with that formula. Turtle Talk with Crush arguably saved The Living Seas pavilion space at EPCOT, and it's executed via a digital puppet operated by a behind-the-scenes cast member doing their best Crush voice.
I sincerely doubt that what makes that experience magical can be replicated with AI in my lifetime. Too much contextual knowledge, too much detail in the nuances of human-human interaction, and too much je ne sais quoi in the timing of getting humor right. I've seen Turtle Talk deal with a particularly excited young person leaning on Crush's "tank" by having Crush look at him and go "Hey little minnow... One of the big humans behind you is gonna come scoop you up. I've seen it happen, lots of times!" You can program that interaction in, but the domain-space of having an interaction for every possible "improv moment" might be outside the bounds of what the next several generations of learning models are capable of.
... or I'm wrong, in which case I look forward to enjoying robo-Seinfeld in the retirement home.
> Most importantly, Olaf can speak and engage in conversations, creating a truly one-of-a-kind experience.
We already live in the world where hackers are pwning refrigerators, I can't wait for prompt injection attacks on animatronic cartoon characters.
> We already live in the world where hackers are pwning refrigerators, I can't wait for prompt injection attacks on animatronic cartoon characters.
It's not necessarily AI controlling the communication. Disney has long had 'puppet' characters whose communication is controlled by a human behind the scenes.
They're already using similar tech for the Mickey meet and greets and the Galaxy's Edge stormtroopers. The details aren't public, but it seems to be a mix of complex dialogue trees with interrupts or context switches, controlled in real time by the actor or operator.
4 replies →
Yep, in this case everything is controlled through a steam deck.
The lack of a video demonstration doesn't really inspire confidence.
there is a detailed video on Disney Research's YT channel: https://youtube.com/watch?v=-L8OFMTteOo
There’s an embedded TikTok showing it off.
Fitting name for a humanoid.
The name Olaf comes from Old Norse Áleifr, combining "anu" (ancestor) and "leifr" (heir/relic), meaning "ancestor's heir" or "ancestor's relic,"
Sometimes the idea of a killer cyborg with a hulking physique and Austrian accent seems absurd. And then we realize the most advanced robots will be made by entertainment companies.
We already have stationary or wheeled/tracked "killer cyborgs" that can easily eeeh terminate anything within their reach and it seems like bipedals are well on their way.
The much greater challenge faced by Disney and Co is making "killer cyborgs" child save and cost effective.
Arguably entertainment requires a much larger range of precision actions that the robot must be able to accomplish, while being in a less controlled environment. That's the cutting edge.
The real reason it won't end up in a park is not the engineering. The problem is the same one as NPCs in computer games: synthetic characters are, to date, just really transparent and boring. The real research question is why.
I guess that's why most computer games don't have NPCs...Oh wait there's entire computer games built entirely around interacting with synthetic NPCs.
There are, of course, limitations to synthetic characters. Even with those limitations there are plenty of entertaining experiences to crafted.
The real challenges are around maintaining and safely operating automous robots around children in a way that isn't too expensive. These constraints place far more limits than those on synthetic characters in video games.
Most people aren't paying 100s or 1000s of dollars to interact with NPCs in video games. If they were, they'd probably expect a lot more and get bored of it quicker.
> The real challenges are around maintaining and safely operating automous robots around children in a way that isn't too expensive.
This is one of the challenges, but only one. The one GP outlined is still very much real - see the Defunctland video on Living Characters for some older examples, but for a recent example, there's the DS-09 droid from Galactic Starcruiser.
Every single non-face character in Disney parks doesn't even talk to guests?
“Prototype-completed design varies.” …Reading this 10 times made me uncomfortably aware of how much I rely on scanning pictures and reading captions to get the gist of an article. A remnant of my academic days perhaps.
I was simply just made uncomfortable by how much CYA the lawyers had to insert into a technical blog post.
This is fun cool tech and I appreciate the insider look, but when the lawyers are peering over your shoulder so much that they need to plaster their "final product may be different" disclaimer even to a r&d audience, well, the Disney Imagineering org sounds more like Disney Legaleering.
I mean, they kind of have to. The last thing you want is some disgruntled Disney-goer trying to get a quick refund or discount for false advertising by saying they, "expected olaf to interact with them, because it looks like he does in the promotional videos"
Cute but I'm more inpressed by the Disney Spiderman stunt robot https://x.com/lukas_m_ziegler/status/1910590914801655814
You can make a robot that's small, soft, and not powerful enough to hurt anyone. Or you can make a robot that's strong enough to carry a laundry basket or climb stairs holding a vacuum cleaner. But you can only operate that big strong robot when there are no humans around. Is that big strong robot an investable idea?
Waymos are kind of big robots that operate with people around.
That is an interesting comparison. More than 600,000 people not in cars are killed on roads every year. If a few hundred thousand were killed by humanoid robots every year it might make a cultural difference.
What do you think the robot makers need to do to have people accept the kind of death count cars deliver?
2 replies →
> an animated character with non-physical movements.
What is a non-physical movement?
Really neat, and made me realize we are getting close to having these type of cute robots at home. With LLMs and voice they would be pretty entertaining companions for many people.
Universal Studios baby dragons did it better.
If you totally alter the character model to fit the envelope of a Boston Dynamics Spot/similar "dog robot', sure.
They can make a two-legged walking robot, but they can't avoid the visible seam in the back of his head?
The tech is amazing, but they need better sewing...
Isn't the robot in the article a prototype?
Yes and hilariously, every single picture of it has a disclaimer stating that the prototype design will vary.
Disney legal is an entity worth studying one day.
Arguably men are two legged walking robots, and men have seams. Even nature couldn't avoid it.
When even Disney can't be bothered to write an article without using the default LLM voice... ugh.
It's a corporate feel that comes from a professional setting and lots of risk aversion. That is exactly what LLMs tend to write, so I sometimes catch myself feeling the "LLM ick" but the article was from before the boom.
So I guess it's just the corporate wash cycle, which I am happy to criticize, LLM generated or not.
Five Nights at Freddys has ruined the joy animatronics for me, they just seem creepy now.
Yeah, I foresee a bite of '27.
This leads me to wonder, when are we likely to have LLMs in robot form in every day life?
You could build one today! Lots of hard problems around a proper humanoid form, but if you're cool with wheels it would be pretty easy to hook up a little robot to GPT.
Look up VLA models; that's essentially plugging the guts of a language model into a transformer that handles joint motion/vision. They get trained on "episodes" i.e. videos from the PoV of a robot doing a task, after training you can ask the model things like: "pick up the red ball and put it into the green cup" etc. Really cool stuff.
https://www.1x.tech/neo, but from what I've heard a lot of times it still has to be remote controlled by a human.
How does a Steam Deck compare to say, TouchOSC on an iPad?
One is a portable linux computer and the other is a MIDI app? What comparison are you hoping for?
I mean't Stream Deck
Strong "Simple Jack" vibes.
Do they wanna build a snowman?
For Paris, I’d honestly be more curious to see a Beast robot from *Beauty and the Beast.
Full-size might be… risky, but a small, friendly mini-Beast could be fun.
[dead]
[dead]
[flagged]
Wrong comments.
Apologies.
For some reason ended up commenting here but should have gone here:
https://news.ycombinator.com/item?id=46345897
Still curious, how Mein Kampf is in the top ten and the included link to a Channel 5 segment ist still relevant also.
But again: apologies for ending up in the wrong spot with this.
>From the way he moves to the way he looks, every gesture and detail is crafted to reflect the Olaf audiences have seen in the film
He looks nothing like a snowman. Snow doesn't look fuzzy. This project appears to focus more on trying to get it moving around in an animated way than getting the character to look right, at least when viewed from photographs.
But how do you know he's made from regular snow and not magical snow that has whatever properties they like? He's literally a talking snowman, lmao.
Watch the movie. Olaf is not fuzzy.