July 31, 2025 • 38 mins
Ready to meet the robots that are about to change your world? In this electrifying episode, we dive deep into the latest breakthroughs in robotics and artificial intelligence, from the debut of cutting-edge humanoid robots like Limx Dynamics’ OLLI and Fourier Intelligence’s GR3, to the rise of robot dogs that can multitask, haul heavy loads, and even perform music!
Discover how affordable, developer-friendly robots are making advanced AI accessible to everyone, and get the inside scoop on specialized machines revolutionizing everything from material sorting to social interaction. We’ll break down the game-changing perception systems like Exhumoid’s Humanoid Occupancy and explore how companies like Skilled are building general-purpose robotics foundation models—ushering in a new era of adaptable, intelligent machines.
Plus, get an exclusive look at the WAIC 2025 exhibition, the world’s hottest showcase for embodied AI, where robots are showing off never-before-seen autonomy, social awareness, and real-world skills. Whether you’re a tech enthusiast, developer, or just curious about the future, this episode is packed with jaw-dropping demos, witty insights, and a few controversial questions about where robotics is headed next.
Don’t miss out on the future—hit play, subscribe, and share this episode with your fellow innovators! The age of intelligent machines is here, and you’ll want to say you saw it first.
Become a supporter of this podcast: https://www.spreaker.com/podcast/tech-threads-sci-tech-future-tech-ai--5976276/support.
You May also Like:
🎁The Perfect Gift App
Find the Perfect Gift in Seconds
⭐Sky Near Me
Live map of stars & planets visible near you
✔Debt Planner App
Your Path to Debt-Free Living
Discover how affordable, developer-friendly robots are making advanced AI accessible to everyone, and get the inside scoop on specialized machines revolutionizing everything from material sorting to social interaction. We’ll break down the game-changing perception systems like Exhumoid’s Humanoid Occupancy and explore how companies like Skilled are building general-purpose robotics foundation models—ushering in a new era of adaptable, intelligent machines.
Plus, get an exclusive look at the WAIC 2025 exhibition, the world’s hottest showcase for embodied AI, where robots are showing off never-before-seen autonomy, social awareness, and real-world skills. Whether you’re a tech enthusiast, developer, or just curious about the future, this episode is packed with jaw-dropping demos, witty insights, and a few controversial questions about where robotics is headed next.
Don’t miss out on the future—hit play, subscribe, and share this episode with your fellow innovators! The age of intelligent machines is here, and you’ll want to say you saw it first.
Become a supporter of this podcast: https://www.spreaker.com/podcast/tech-threads-sci-tech-future-tech-ai--5976276/support.
You May also Like:
🎁The Perfect Gift App
Find the Perfect Gift in Seconds
⭐Sky Near Me
Live map of stars & planets visible near you
✔Debt Planner App
Your Path to Debt-Free Living
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Imagine stepping into a world where the future isn't just
you know, something on the horizon, it's already here, like
unfolding right now, right in front of us. Uh huh.
Picture advanced machines seamlessly serving you ice cream at an expo,
or effortlessly juggling your groceries right and even incredibly conducting
(00:21):
a live symphony orchestra with actual nuance and grace.
Speaker 2 (00:25):
It sounds like science fiction, doesn't it totally?
Speaker 1 (00:28):
And what if the most thrilling feats of say, athleticism,
weren't just about human skill anymore, but about the astonishing
agility of these intelligent machines.
Speaker 2 (00:39):
It's a huge shift.
Speaker 1 (00:40):
We're i mean, really living through a truly transformative time,
a moment where that line between what humans can do
and what machine intelligence can do is blurring faster than well,
faster than most of us can probably wrap our heads around.
Speaker 2 (00:52):
Absolutely, and it makes you wonder.
Speaker 1 (00:54):
Exactly, how are these machines doing such incredible things, things
we couldn't have imagined just a few years ago. What
are the real brain through is driving this just breathtaking pace?
Speaker 2 (01:03):
That's the core question.
Speaker 1 (01:04):
So today we're doing a deep dive We're looking at
some truly groundbreaking reports, some new insights that really show
the incredible speed of innovation in AI robotics, and this
deep dive it's designed for you, our listener, to get
through all the noise, pull out the most important, maybe
surprising bits.
Speaker 2 (01:25):
Of knowledge, and give it some context.
Speaker 1 (01:27):
Exactly with engaging context. Think of it like your shortcut
to being genuinely well informed on a topic that is,
let's face it, reshaping everything.
Speaker 2 (01:36):
It really is.
Speaker 1 (01:37):
Our mission here is pretty broad. We'll explore everything from
nimble humanoids doing daily tasks with surprising finesse.
Speaker 2 (01:44):
Yeah, the dexterity is getting amazing.
Speaker 1 (01:46):
Two powerful rugged robot dogs tackling complex, even dangerous terrains,
and crucially, the brain's behind it.
Speaker 2 (01:54):
All, the AI system's powering them. That's key.
Speaker 1 (01:57):
We'll unpack how robots aren't just evolving step by step
a truly revolutionizing things around us.
Speaker 2 (02:01):
It feels like an exponential curve.
Speaker 1 (02:03):
Sometimes right from making advanced robotics way more accessible, more
affordable than ever.
Speaker 2 (02:07):
Which is huge for innovation.
Speaker 1 (02:09):
To developing these unified brains that promise this universal intelligence
powering any robot for any task. That sounds wild, it's
a massive ambition and then showcasing just stunning real world
capabilities that honestly, they're going to make your jaw drop.
They really challenge your idea of what robots could even do.
Speaker 2 (02:30):
And what's truly fascinating here isn't just what they can
do now, but the how you know, the underlying tech. Yeah,
the mechanics, the software, yeah, exactly, the sophisticated control systems,
the multimodal perception letting them understand the world in a
much richer way, not just see it.
Speaker 1 (02:48):
Okay, multimodal perception. We'll get into that.
Speaker 2 (02:50):
And also the strategic shifts. Companies are making big, sometimes
audacious moves to get these machines out of the labs
and you know, into our everyday.
Speaker 1 (02:58):
Lives, right out into the real world.
Speaker 2 (03:00):
And this rapid progress. It brings up a really important
question for all of us, doesn't it. What does all
this mean for how we'll live, how will work, how
we'll interact with technology in the well, the very near future.
Speaker 1 (03:10):
It's not abstract anymore.
Speaker 2 (03:12):
Not at all. It's a question we need to be
thinking about because these intelligent machines are moving fast from
theory and lab tests to actual tangible things impacting our
lives daily.
Speaker 1 (03:24):
Okay, let's dive into the first big piece of news, then,
the one that's got the robotics world buzzing. That's the
official launch of Limix Dynamics new humanoid robot Oli ah.
Speaker 2 (03:34):
Yes, Oli formerly CL three that used to.
Speaker 1 (03:37):
Be CL three, And it's making waves not just for
what it can do, but maybe even more for its
price tag, which kind of stunning really is. We're taking
a starting price around one hundred and fifty eight thousand,
which translates to roughly twenty two thousand US dollars.
Speaker 2 (03:52):
Twenty two thousand for a humanoid.
Speaker 1 (03:54):
That's not just a number, right, that feels like a statement. Yeah,
So for those of us watching this field accelerate, what
is that price point actually mean for the market for accessibility?
Speaker 2 (04:03):
Well, if you connect this to the bigger picture, it's
genuinely significant. I'd even say transformative for the industry.
Speaker 1 (04:08):
Ysorm of hew think about it.
Speaker 2 (04:10):
For a fully capable humanoid robot. It stands five foot
five styles so human.
Speaker 1 (04:14):
Size basically fits in human environments.
Speaker 2 (04:17):
Exactly, weighs fifty five kilos, which is manageable, and it
has thirty one degrees of freedom.
Speaker 1 (04:22):
Okay, thirty one degrees of freedom? Break that down quickly.
What does that allow?
Speaker 2 (04:25):
It means an extraordinary range of motion. Think about your arm,
maybe seven degrees of freedom. Lets you do all sorts
of things, thirty one across the whole body. That enables
incredibly complex, nuanced, almost human like movements, fine motor control,
navigating cluttered spaces gracefully.
Speaker 1 (04:42):
Got it so, highly capable.
Speaker 2 (04:45):
Highly capable, and at that price point twenty two thousand dollars,
that positions o ALI as a truly affordable option.
Speaker 1 (04:52):
Affordable compared to what typically what were we looking at?
Speaker 2 (04:55):
Multiples of that hundreds of thousands easily for research grade
humanoids with similar specs, often millions for the really high
end ones.
Speaker 1 (05:02):
Wow.
Speaker 2 (05:03):
Okay, So this affordability is the key. It means OLILI
is accessible not just to the big tech giants with
massive R and D budget.
Speaker 1 (05:10):
Right at the usual suspect.
Speaker 2 (05:12):
But crucially to researchers, innovative startups, universities all over the world.
It effectively democratizes access developing on advanced humanoid platforms.
Speaker 1 (05:20):
So it spreads the innovation potential around precisely.
Speaker 2 (05:23):
We're seeing these incredible machines move beyond being exclusive toys
for a few big players. It opens them up to
a much broader base of innovators who are you know,
really eager to push the boundaries now that they can
actually get their hands on the hardware.
Speaker 1 (05:37):
That is genuinely remarkable. But when you hear fully capable
at such a low price twenty two k and does
make you wonder how do they manage that? What are
the technical choices that let them hit that price without
you know, sacrificing core abilities.
Speaker 2 (05:53):
That's a great question, And one key thing attestament to
some clever engineering for cost efficiency is OBII in house
developed six axis.
Speaker 1 (06:02):
IMU im you inertial measurement units exactly.
Speaker 2 (06:05):
Now. IMU sounds super technical, I know, but it's absolutely crucial.
It's like the robot's inner ear, especially for a bipedal
one that needs to balance.
Speaker 1 (06:13):
Like how we stay upright without thinking about.
Speaker 2 (06:15):
It pretty much, imagine trying to walk on two legs
if you didn't constantly know how your body is tilted,
how fast it's turning or accelerating. The IMU provides that
precise real time data orientation, angular velocity, linear acceleration.
Speaker 1 (06:28):
All constantly feeding into the control system constantly.
Speaker 2 (06:31):
It's vital for precise motion control and most importantly for
dynamic balance. For all at LIE to walk smoothly stably,
especially in human spaces, it needs to constantly monitor and
instantly adjust.
Speaker 1 (06:44):
So developing that sophisticated piece in house was probably a
big cost saver while keeping performance high.
Speaker 2 (06:51):
Very likely, it gives them direct control over a core component,
ensuring it meets their needs for precise movement, without relying
on expensive off the shelf parts that might have features
they don't need or not integrate as well.
Speaker 1 (07:04):
Okay, makes sense. And what about for the people who
want to actually use all I, the developers, the researchers,
the students. How easy is it for them to get started?
Speaker 2 (07:12):
This is where Limus Dynamics really seems to be playing
it smart, showing a very forward thinking strategy. ALLI comes
with a modular.
Speaker 1 (07:19):
SDK software development kit right.
Speaker 2 (07:20):
And a fully open interface, and it allows for Python
based customization.
Speaker 1 (07:24):
And control Python. Okay, that's huge.
Speaker 2 (07:26):
It's a game changer for the developer community. It means
they aren't locked into some proprietary complex system. They can
program OLI using Python, which is one of the most popular,
versatile developer friendly languages out there, so.
Speaker 1 (07:40):
The learning curve is lower, the potential developer pool is massive.
Speaker 2 (07:44):
Exactly. It supports rapid iteration. Developers can quickly test new ideas,
experiment with different applications, integrate OLI into diverse scenarios from
fundamental AI research to practical service robotics.
Speaker 1 (07:57):
So limics isn't just selling a robot. They're selling a platform.
Speaker 2 (08:00):
That seems to be the strategy, a development platform, almost
like an advanced computer you can program for tons of
different tasks, not just a fixed function machine. This openness
plus the price, plus the real time control, it makes
OLI a serious disruptor, especially.
Speaker 1 (08:16):
For anyone looking for that balance between high capability and
genuine accessibility. Right.
Speaker 2 (08:21):
It kind of invites you, the listener to think, Okay,
if I had access to this, what could I build?
Speaker 1 (08:25):
It?
Speaker 2 (08:25):
Fosters innomation way beyond Limits's own labs.
Speaker 1 (08:28):
Okay, so OILI is out there making waves with affordability.
But then there's the anticipation game. Fourier Intelligence just dropped
a sneak peak of its upcoming GR three humanoid.
Speaker 2 (08:39):
Ah the GO three teaser. Yeah, that got people talking.
Speaker 1 (08:42):
It's just a digital render, not the real thing yet.
But enough to create buzz ahead of the full reveal
on August sixth. What's the strategy here? Why release just
a render?
Speaker 2 (08:53):
Well, what's fascinating is the timing and the subtle hints
in that render. They did show a prototype briefly at
WAIC twenty twenty five in Shanghai.
Speaker 1 (09:03):
The event will definitely circle back to for.
Speaker 2 (09:04):
Sure, But this new render, it's suggest something sleeker, more
advanced and evolution, moving maybe from just function towards a
more refined design.
Speaker 1 (09:13):
So what hints are we seeing? What is refined suggest the.
Speaker 2 (09:15):
Focus seems to be on well, refine mobility, smoother, more
precise movements, better agility, also enhance mechanical balance, which implies
more stability, especially when things get unpredictable, and potentially more
sophisticated AI interaction capabilities.
Speaker 1 (09:29):
Meaning it could understand and respond.
Speaker 2 (09:31):
Better, possibly respond more intelligently, more intuitively to its environment,
maybe to human commands. We don't know for sure yet,
but that's the vibe.
Speaker 1 (09:40):
And releasing this teaser just two weeks before launch pretty
standard hype building.
Speaker 2 (09:46):
It's smart marketing in a very competitive field. It puts
Fourier intelligence right back in the conversation in this rapidly
accelerating humanoid race. You know, companies in China, globally, everyone's
pushing hard for real world cap robots at scale.
Speaker 1 (10:01):
So it's about signaling leadership, attracting attention.
Speaker 2 (10:04):
Exactly, building anticipation, signaling they're a major player, attracting potential partners,
creating a narrative. If this renders anything to go by,
the g R three could be a pretty significant step
for Fouria, maybe their most important humanoid yet.
Speaker 1 (10:18):
Okay, so from anticipation to something we've actually seen in action,
let's talk about Droidip's Walker Too. It was already kind
of a fan favorite, right, this bipedal roma.
Speaker 2 (10:26):
Yeah, Walker two has definitely had a following.
Speaker 1 (10:28):
It just got a major upgrade, Yeah, autonomous navigation. But
the surprise was they also introduced a new companion, Rina,
the wheeled one, right, and together they put on this, well,
this synchronized show. Beyond just looking cool, what does that
performance tell us about where multi robot collaboration is heading.
Speaker 2 (10:48):
This brings up a really interesting point about how robots
are moving beyond just doing their own thing towards expressive
and really intricate coordination like a team. Exactly, Walker Too
can how navigate dynamic places all by itself. It responds
in real time, keeps its balance, understands the space around it.
It's not just following lines on the.
Speaker 1 (11:08):
Floor, it's perceiving and deciding right.
Speaker 2 (11:11):
And Rena the wheeled one. She's built for speed and
smooth movement on flat surfaces, very agile, so different strengths,
totally different strengths, and their duet style dance routines at
WWAC twenty twenty five weren't just for show. They really
showcased Droydep's mastery in getting different types of robots to collaborate.
Speaker 1 (11:27):
Two different locomotion systems working together precisely.
Speaker 2 (11:30):
Imagine that a bipedal robot good for human spaces and
a wheeled one good for speed, performing a complex synchronized routine,
reacting to each other, reacting to the music, the environment seamlessly.
Speaker 1 (11:43):
That coordination must be incredibly complex to program.
Speaker 2 (11:46):
Oh. Absolutely. The contrast between Walker two sort of bipedal
elegance and Rena's smooth rolling agility really highlighted their synergy.
It showed how different robot types can complement each other.
Speaker 1 (11:58):
And the implications go beyond dance through teams. Obviously, for sure.
Speaker 2 (12:01):
Think about complex operations in the future, logistics and a
busy warehouse, maybe disaster response. It probably won't be one
single robot doing everything.
Speaker 1 (12:10):
It'll be teams, heterogeneous teams.
Speaker 2 (12:13):
Exactly specialized machines working together, some agile, some strong, some fast,
optimizing the whole system. This demo hnts that future where
robots coordinate complex tasks in industry, maybe in service roles,
or even become more engaging companions at home.
Speaker 1 (12:28):
Okay, shifting gears a bit. One clip that really went
viral from WAIC twenty twenty five involved Unitre's g one
humanoid ah.
Speaker 2 (12:37):
The fighting robot demo. Yeah, that blew up online.
Speaker 1 (12:39):
It sounds like pure sci fi. The g One squaring
off against professional fighters. You'd think it's a mismatch, but
apparently the g one held its own.
Speaker 2 (12:47):
Held its own is putting it mildly. What was fascinating
was the sheer, physical prowess, the agility the g One displayed.
It has twenty six high performance.
Speaker 1 (12:56):
Joints, which is a lot of control points, a lot.
Speaker 2 (12:59):
And it demonstrated this martial arts style agility that honestly
stunned people in the videos you see it blocking ducking,
striking back with shockingly fluid motions like it knew it
was coming, almost as if it was reading the opponent's
moves in real time. And these were professional fighters, not actors.
They were going at it, but the g One kept
its robotic composure, didn't fall over, didn't.
Speaker 1 (13:21):
Break, so definitely more than just a fun stunt.
Speaker 2 (13:23):
Way more, it highlighted how far humanoid robotics has come
in motion control, in reflex training. The g One wasn't
just walking or doing preset moves. It was sparring with
trained athletes, dynamic interaction, physical resilience. Stuff that was pure science.
Speaker 1 (13:38):
Fiction very recently, and it sparked debates right.
Speaker 2 (13:40):
Oh yeah, across tech circles, even in combat sports. People
were asking, are we seeing the dawn of robotic fighters
or is this mainly just a very very impressive demo
of motion control and reaction speed.
Speaker 1 (13:51):
Either way, it shows a huge leap in capability.
Speaker 2 (13:54):
A full blown tech spectacle, absolutely yeah, proving robots can
do these high speed hidexs starity things with surprising control
and toughness.
Speaker 1 (14:03):
When you say shocking fluidity and reading moves, that sounds
like incredibly advanced AI and control behind the scenes. What
kind of tech makes that possible?
Speaker 2 (14:13):
Yeah, to get that kind of fluidity and real time reaction,
you're looking at extremely advanced, low latency control systems, almost
certainly using sophisticated AI, maybe real time reinforcement learning or
predictive modeling.
Speaker 1 (14:27):
So it's not just reacting, it's anticipating.
Speaker 2 (14:29):
Potentially, it needs incredibly fast sensor feedback, high speed cameras,
probably force and torque sensors in the joints and limbs.
The AI has to process all that data instantly to
generate the right countermove. It's not just speed, it's intelligent
adaptive response.
Speaker 1 (14:44):
And the fact that it didn't you fall apart, right.
Speaker 2 (14:46):
That implies a really robust mechanical design, too able to
take impacts and efficient power management to keep going during
that intense activity.
Speaker 1 (14:54):
So beyond the wow factor of fighting robots, where else
could this kind of agility be used? What other high
dexterity applications could this unlock?
Speaker 2 (15:03):
Oh? Tons of places? Imagine robots with this agility doing
complex assembly and high tech manufacturing, placing tiny components with
split second precision.
Speaker 1 (15:13):
Faster and maybe more accurate than.
Speaker 2 (15:15):
Humans potentially, or think about emergency response robots navigating rubble,
unstable buildings, clearing debris, maybe even performing rescues in situations
too dangerous.
Speaker 1 (15:26):
For people, hazardous environments too, like nuclear sites or chemical handling.
Speaker 2 (15:30):
Exactly performing intricate tasks with precision, endurance and safety that
humans just can't match. This level of agility, precision, resilience.
It could fundamentally change fields needing fine motor skills, fast reactions,
and tough physical performance, logistics, healthcare construction. The list goes on.
Speaker 1 (15:49):
Okay, so we've covered some amazing physical feats, agility, coordination, resilience,
but the body is only half the story, maybe less.
The real revolution is often in the brain.
Speaker 2 (15:59):
The intelligence driving it all.
Speaker 1 (16:00):
Absolutely, So let's shift our deep dive there to the intelligence,
the perception, the big ambitions to create unified AI for
all robots. After staying quiet for months, Skilled finally gave
a glimpse into their huge project.
Speaker 2 (16:16):
Yeah. Skilled emerged from Stealth last year, and now we're
hearing more.
Speaker 1 (16:20):
They're trying to build a single general purpose robotics foundation
model one AI to power any robot for any task.
Speaker 2 (16:29):
That sounds monumental, almost sci fi. Why does this Skilled brain.
Speaker 1 (16:34):
Well connecting it to the bigger picture. It's a truly
audacious vision, potentially revolutionary for robotics. The skilled brain is
basically a unified control system designed to operate different kinds of.
Speaker 2 (16:45):
Robots, humanoids, quadrupeds.
Speaker 1 (16:47):
Arms exactly humanoids likeli those quadrupeds we talked about, even
simpler robotic arms, all using the same underlying AI architecture.
Speaker 2 (16:55):
How is that even possible? Usually they need separate programming.
Speaker 1 (16:57):
Right, that's a traditional way, but skilled different think about
how we learn. We learn general principles about picking things up, physics,
object properties, how to grip, and we apply that whether
we're picking up a feather or a brick with our
hands or a tool. So they want robots to have
that same kind of general understanding.
Speaker 2 (17:14):
That's the goal, generalized intelligence for robots, and they're doing
it by training their model on massive, incredibly diverse data.
Speaker 1 (17:22):
Sets of actions, well kind of data.
Speaker 2 (17:24):
Everything, huge amounts of Internet videos showing humans and machines
doing things in all sorts of contexts, high fidelity simulation
data where they can learn safely and quickly in virtual worlds,
and tons of real world sensor data from actual robots
doing actual tasks.
Speaker 1 (17:40):
So quantity and variety of data is key.
Speaker 2 (17:42):
Absolutely crucial. The goal is to break down those traditional
silos in robot control where you train one model for walking,
another for grasping, another for a.
Speaker 1 (17:50):
Different robot, which is slow and expensive.
Speaker 2 (17:53):
Incredibly slow and expensive.
Speaker 1 (17:54):
Yeah.
Speaker 2 (17:55):
Skilled wants a scalable, adaptable intelligence that generalizes across different
robot bodies, a single brain that could theoretically run a
warehouse robot one minute and a home assistant the next.
Speaker 1 (18:05):
Which raises the question could this really speed up robot
development everywhere?
Speaker 2 (18:10):
That's the billion dollar question, isn't it. If they succeed,
it could massively accelerate things, making it faster and cheaper
to deploy smart robots into completely new areas. They're aiming
to build the universal intelligence to run them all. It
could fundamentally reshape the entire.
Speaker 1 (18:26):
Field that lead from bespoke AI for each robot to
a universal model. Why is breaking silos so critical for
robotics and what makes Skill think they can pull off
such a huge challenge.
Speaker 2 (18:38):
Well, the traditional way, being so fragmented, is a huge
bottleneck developing separate models for every robot every task. It's
resource intensive, slow, hard to scale. Imagine needing a unique
brain for every single thing you want a robot.
Speaker 1 (18:52):
To do, progress would crawl exactly.
Speaker 2 (18:55):
A unified foundation model, though, learns general principles of interacting
with the physical world from watching countless examples. It's these
patterns generalizes skills, kind of like how large language models
llms revolutionized language.
Speaker 1 (19:07):
Tasks like GBT for language right.
Speaker 2 (19:09):
Instead of training separate models for translation, summarization Q and
A one big foundational LLM can adapt skilled. Is trying
to do the same for physical actions, building essentially a
GPT for robots learning the grammar and physics of interaction
from massive.
Speaker 1 (19:25):
Data so the data is absolutely fundamental.
Speaker 2 (19:28):
Data is king is an understatement here. Training on those massive,
diverse action data sets is how they give the AI
a comprehensive, nuanced understanding of how things work. Objects forces environments.
Speaker 1 (19:40):
Like learning to open any door, not just one specific.
Speaker 2 (19:43):
Door, precisely seeing thousands of examples different doors handles people
opening them. The physics involved that breadth and variety are
crucial for robots to be truly robust and adaptable and messy,
unpredictable real world situations and skilled itself.
Speaker 1 (19:57):
What's their background?
Speaker 2 (19:58):
What makes them compelling is the team founded in twenty
twenty three by really distinguished professors from Carnegie Mellon.
Speaker 1 (20:04):
Universities CMU, a powerhouse in robotics.
Speaker 2 (20:07):
Research, absolutely and backed by robotics pioneers who've done some
of the most important work in the field over the
last decade. This isn't just a startup with a flashy idea.
It's deep tech with a very strong academic and research foundation.
Speaker 1 (20:21):
And now they're shifting from research to application.
Speaker 2 (20:24):
Exactly taking that deep knowledge and applying it to this
massive real world problem. They're not just building robots, they're
aiming to build the universal brain for robots, the foundational
intelligence that could unlock a whole new generation of capabilities.
Speaker 1 (20:40):
Okay, so that's the brain. Now, how do robots understand
the world To feed that brain?
Speaker 2 (20:46):
We usually think cameras, right, vision visions is definitely dominant historically.
Speaker 1 (20:50):
Meximoyd just unveiled something called humanoid Occupancy. They're calling it
the world's first general purpose multimodal perception system specifically for humanoids.
Speaker 2 (21:00):
That sounds like more than just cameras.
Speaker 1 (21:02):
It is much more, and what's fascinating is how it
tries to mirror or even extend human perception you're right.
Traditional systems rely heavily on cameras.
Speaker 2 (21:10):
Robots mostly see, but we use more than just our eyes,
way more we use site sure, but also hearing, detecting sounds,
figuring out where they're coming from, what they mean, touch, feeling, textures, forces, temperature,
and our spatial awareness appropriateception knowing where our body parts
are without looking.
Speaker 1 (21:29):
Okay, so humanoid occupancy tries to give robots all of that.
Speaker 2 (21:33):
That's the idea, integrating audio, tactile, touch, and spatial sensing
into one unified system. And it's not just about adding
more sensors. The key is sensor.
Speaker 1 (21:43):
Fusion, combining the data intelligently.
Speaker 2 (21:45):
Intelligently combining and interpreting data from all these different sources visions, sound, touch,
spatial awareness in real time to build a truly holistic,
robust internal picture of the environment.
Speaker 1 (21:57):
Can you give an example, how does that help?
Speaker 2 (21:58):
Sure? Imagine a row in a busy factory. Its camera
view might be blocked by a stack of boxes for
a second.
Speaker 1 (22:04):
A vision only robot would be blind, right.
Speaker 2 (22:07):
But a multimodal robot might hear the distinct sound of
an approaching forklift. Its tactile sensors might feel a slight
vibration in the floor as the heavy vehicle gets closer.
Speaker 1 (22:17):
So even without seeing it, it knows something's coming exactly.
Speaker 2 (22:20):
By fusing these different inputs, it gets a much richer,
more human like understanding. It can make safer, more nuanced
decisions than a robot just relying on site.
Speaker 1 (22:30):
This sounds like the foundation for well more intelligent behavior.
Speaker 2 (22:34):
It really is. It lays the groundwork for a new
era of embodied intelligence, robots moving from just reacting to
what they see to making context to were decisions precisely
based on a fuller understanding of the world, whether they're
in homes, factories, outlowers. It's a huge shift from reactive
automation just doing pre programmed things or reacting to visual cues,
(22:55):
to proactive intelligent assistance.
Speaker 1 (22:57):
Where they can anticipate, adapt make better.
Speaker 2 (23:00):
Judgment yes, which dramatically improves their safety, their efficiency, their adaptability,
especially in unpredictable real world settings. It's a profound step up.
Speaker 1 (23:09):
All these individual breakthroughs, affordable hardware, unified brains, multimodal perception,
they really seemed to come together at one big event
WAIC twenty twenty five in Shane High.
Speaker 2 (23:20):
Oh Yeah, WIC was huge this year, a real landmark.
Speaker 1 (23:23):
Described as the largest embodied AI and robotics showcase ever.
It painted this vivid picture of the future actually happening.
What really stood out there? How did it show the
convergence of everything we've been talking about.
Speaker 2 (23:36):
WAIC twenty twenty five was honestly a powerful statement. Embodied
AI isn't niche anymore. It's taking center stage globally. The
scale was just breathtaking.
Speaker 1 (23:47):
Give us some numbers.
Speaker 2 (23:48):
Over three thousand AI innovations showcased, more than one hundred
and fifty different humanoid robots and over sixty brand new
robot models debuting all in one place. It felt like
a turning point.
Speaker 1 (24:00):
And the robots themselves what was different.
Speaker 2 (24:02):
They weren't just slightly more capable. They were noticeably smarter, faster,
and interestingly more socially aware.
Speaker 1 (24:08):
Socially aware how so.
Speaker 2 (24:09):
Meaning they could interact with humans and navigate crowded spaces
with more nuanced, more understanding. They showed greater autonomy, adapted
better to new situations, handled complex environments more easily. Interactions
felt more fluid, more intuitive.
Speaker 1 (24:22):
So the theme was really about readiness for the real world.
Speaker 2 (24:25):
Absolutely, the overarching themes were autonomy, adaptability, and real world integration.
This wasn't just a show of lab prototypes that might
be useful someday. It felt like a showcase of machines
ready or very nearly ready to operate among us in
our daily lives.
Speaker 1 (24:42):
And this cannakes back to things like skilled unified brain
and Eximoid's multimodal perception.
Speaker 2 (24:47):
Directly, those concepts a universal brain, rich multisensory understanding aren't
just features. There are critical enablers for the kinds of autonomous, adaptable,
socially aware robots we saw everywhere at WAIC.
Speaker 1 (25:00):
WAIC proved the pieces are coming together.
Speaker 2 (25:02):
It vividly demonstrated that the next wave is about machines
truly being integrated into our environments, doing diverse tasks intelligently, fluidly, independently.
The sheer scale and variety of innovation it signaled a
field that's mature and accelerating incredibly fast. We're on the
cusp of a major robotic revolution impacting well everything.
Speaker 1 (25:21):
Let's zoom in on that real world impact. At WAIC
we saw robot stepping directly into service and retail rolls,
moving beyond factories. Tell us about the Pepsi cool bot
and that robotic mini store concept.
Speaker 2 (25:32):
Right, those were fascinating because they showed immediate, tangible practicality
robots moving into public consumer facing jobs.
Speaker 1 (25:40):
So the Pepsi cool bot, what was that.
Speaker 2 (25:42):
It was a collaboration between a Gibbet and Pepsi. And
this robot wasn't just static. It autonomously navigated the busy
expo floor, weaving through crowds, yeah, with impressive precision. It
even had a certain charm in its movements, apparently delivering
drinks directly to people.
Speaker 1 (25:58):
So not just a gimmick, it didn't.
Speaker 2 (25:59):
Feel like one, and it felt like a clear glimpse
into the future of retail automation, handling, distribution, inventory, even
customer interaction in a dynamic public space.
Speaker 1 (26:09):
And the robotic mini store that.
Speaker 2 (26:11):
Was from Gelbit Robotics basically a fully autonomous retail kiosk,
a self contained little store. Visitors could walk up, interact
with it, browse products, choose something, and complete the purchase,
all without any human staff.
Speaker 1 (26:25):
Wow, like a really advanced vending machine.
Speaker 2 (26:28):
Kind of, but much more interactive and flexible. Again, these
aren't just concepts. There are operational previews of a more
automated efficient, maybe more convenient retail landscape.
Speaker 1 (26:38):
What are the implications here for businesses and consumers?
Speaker 2 (26:41):
For businesses, potential for optimized operations, lower staffing costs maybe
two hundred and four to seven. Service For consumers maybe
instant frictionless shopping, more access to goods. It points to
a future where convenience and efficiency are increasingly driven by
smart automation. It changes how business is run and how
we shop.
Speaker 1 (27:02):
Okay, Moving from retail to the tougher environments factories, warehouses,
WAIC showed big strides there too. Leisure Robotics had new
materials sorting tech.
Speaker 2 (27:13):
Yes, Lijiu showed a major upgrade. Their humanoids can now recognize, classify,
and organize a wide variety of materials in.
Speaker 1 (27:21):
Real time, beyond just color and shape.
Speaker 2 (27:23):
Way beyond. In the demos, the robot was accurately distinguishing
between plastics, different metals, fabrics, glass, sorting them into bins flawlessly.
Speaker 1 (27:30):
Does it do that?
Speaker 2 (27:31):
It uses advanced visual recognition, but also AI driven texture analysis,
maybe other sensors, and contextual reasoning to figure out what
it's holding.
Speaker 1 (27:39):
And why is that important? Material sorting?
Speaker 2 (27:41):
It's a hugely tedious repetitive, often dirty, and sometimes hazardous
job for humans, especially in recycling plants or waste management.
Lijiu's tech shows AI turning humanoids from show pieces into
practical helpers for these tough tasks, reduces human burden boost
sufficiency could transform large scale res cycling, maybe even home
waste separation down the line.
Speaker 1 (28:03):
Then there was AI, two robotics and their humanoid Aibo
doing palatizing.
Speaker 2 (28:09):
Aibo showcased whole domain full body palatizing, and it was
seriously impressive.
Speaker 1 (28:14):
What does full body palatizing mean?
Speaker 2 (28:16):
It means the robot used its entire body torso rotation,
core strength, fine finger control to lift and stack boxes
onto a palette super efficiently in under sixty seconds. It
stacked multiple boxes perfectly.
Speaker 1 (28:28):
How did it manage that coordination?
Speaker 2 (28:30):
It dynamically adjusted its posture, anticipated how the boxes would shift,
fine tuned its grip in real time, all autonomously using
its visual sensors and AI coordination. No resets, no hesitation,
no dropped boxes, just flawless execution.
Speaker 1 (28:43):
That sounds like a game changer for logistics.
Speaker 2 (28:45):
Absolutely, palatizing is physically demanding, high risk for human injury.
Aibo showed robots can do it faster, more consistently, maybe safer.
It's a clear preview of the next gen factory or
distribution center and.
Speaker 1 (28:58):
Chinten Cyborg roboticslaunched their ro one humanoid.
Speaker 2 (29:01):
Yes, the Cyborg R one made its official debut. Sleek
industrial design but clearly built for real world function.
Speaker 1 (29:09):
What could it too?
Speaker 2 (29:09):
It showed smooth walking, stable balancing, precise hand control, using tools,
transferring objects, even responding to gestures naturally.
Speaker 1 (29:18):
What makes it adaptable?
Speaker 2 (29:19):
It has a modular architecture and advanced AI high performance joints,
good three D spatial awareness and natural language processing so
it can understand complex voice commands.
Speaker 1 (29:28):
So it could be used for lots of different factory
or service jobs.
Speaker 2 (29:31):
That's the idea multipurpose assistant. Its plug and play flexibility
makes it adaptable for different tasks. This whole suite of
industrial robots from le Zeu AI two Cyborg it signals
a real acceleration towards useful task specific AI turning humanoids
into indispensable real world workers.
Speaker 1 (29:49):
Humanoids get a lot of attention, but the robot dogs
the quadrupeds are also making huge leaps right, especially in tough.
Speaker 2 (29:55):
Spots definitely quadruped progress is incredibly impressive and arguably more
practical in many near term scenarios.
Speaker 1 (30:02):
We saw Rambo from ETH Zurich doing multitasking.
Speaker 2 (30:07):
Yeah. Rambo was amazing, developed by gin change In's team
at ETH. It's stunned people by pushing a shopping cart
while balancing a plate.
Speaker 1 (30:13):
Brunch hustling robot Boss.
Speaker 2 (30:15):
Exactly look charming effortless, but underneath is incredibly sophisticated control.
Rambo stands for RL augmented model based whole body control
for loco manipulation.
Speaker 1 (30:27):
Okay, unpack that acronym RL augmented.
Speaker 2 (30:30):
It uses a hybrid approach precise model based planning for
accuracy combined with real time reinforcement learning feedback.
Speaker 1 (30:36):
So it learns and adapts on the fly right.
Speaker 2 (30:38):
Which makes it super robust and adaptable in dynamic situations.
It lets Rambo seamlessly combine walking and manipulation, which is
a huge challenge.
Speaker 1 (30:46):
In robotics, handling complex coordination naturally.
Speaker 2 (30:49):
Yeah, adjusting to uneven floors, recovering from bumps, balancing things
while moving. It showed amazing fluidity and resilience, a big
leap in mobile manipulation and.
Speaker 1 (31:00):
Deep Robotics showed a whole range of quadrupeds.
Speaker 2 (31:03):
They did a real showstopper, from smaller ones for tight
spaces to heavy duty industrial models, impressive power agility, real
world functionality. So was their focus embodied AI. These robots
weren't just following commands, They were thinking, using perception, real
time decision making, adaptive learning, what kind of tax, navigating
(31:24):
complex obstacles, climbing stairs, easily carrying heavy payloads over rough ground,
responding to voice and gestures intuitively. They showed uses from
power inspections and emergency response to logistics and construction, proving
their AI platform is scalable and versatile.
Speaker 1 (31:39):
They seem ready for serious work.
Speaker 2 (31:41):
Absolutely Deep Robotics basically brought the future to the show floor,
proving quadrupeds with embodied AI are already solving real problems
with speed, accuracy, and autonomy.
Speaker 1 (31:50):
And then there was the heavy lifter phi In Titan
from the Southern Power Grid.
Speaker 2 (31:55):
Yes, py On Titan is breaking records designed specifically for
power grid inspection. It tackles the load bearing limits of
older robots.
Speaker 1 (32:04):
How does it manage heavy loads and tough terrain.
Speaker 2 (32:06):
It has an innovative hybrid mobility system wheels and legs.
Speaker 1 (32:10):
Best of both worlds, kind.
Speaker 2 (32:12):
Of unmatched adaptability across gravel slopes, industrial debris places critical
for inspections but often dangerous or inaccessible for humans or
wheeled robots.
Speaker 1 (32:22):
And it carries tools exactly.
Speaker 2 (32:24):
It's a platform for specialized gear HD, thermal cameras, lightar,
for three D mapping, robotic arms, other sensors. It excels
at fighting defects, reading meters, autonomous.
Speaker 1 (32:35):
Patrolling with AI helping identify problems.
Speaker 2 (32:37):
Yes, Its AI recognition systems spot abnormalities in real time,
reduces manual labor, boosts safety dramatically and high voltage areas.
It's a major leap for smart grid maintenance and autonomous
robotics and critical infrastructure, transforming how power inspections get done.
Speaker 1 (32:52):
Okay, from heavy industry to something completely different. Astrobots s
one robot performing with the Beijing National Orchestra as a
conductor and music.
Speaker 2 (33:00):
Yeah, that one truly blurred the lines automation meets artistry.
It sounds completely sci fi, doesn't it?
Speaker 1 (33:07):
A robot conductor? Totally unpredictable? What made it so jaw dropping?
What did it actually do?
Speaker 2 (33:12):
It raises fascinating questions about performance, collaboration, even emotion in
human robot interaction. The S one guided the orchestra with
remarkably accurate, fluid conducting.
Speaker 1 (33:23):
Gestures, setting tempo dynamics, queuing sections, all in real time.
Speaker 2 (33:27):
And it wasn't just rigid, pre programmed movements. It was dynamic,
responsive to the live musicians.
Speaker 1 (33:33):
It adapted to the orchestra.
Speaker 2 (33:34):
That's what was so impressive. It seemed to genuinely interact,
responding to subtle musical shifts. It even acknowledged the audience.
People felt a surprisingly emotional connection, a real sense of presence.
Speaker 1 (33:45):
Wow. Not just robotic repetition.
Speaker 2 (33:47):
Not at all. It was a profound proof of concept
showing humanoids can engage in high level collaboration, meeting rhythm timing,
nuanced human interaction, from smooth wrist flicks conveying musical ideas
to keeping the whole orchestra together. It blurred that line
between automation and artistry beautifully.
Speaker 1 (34:05):
So robots and cultural experiences.
Speaker 2 (34:07):
It marks a new chapter, robots not just working behind
the scenes, but becoming integral, maybe even inspiring parts of
culture and art, showing their learning to move, feel, perform
like us suggests the future of robots as genuine collaborators
in creative fields. It really expands your idea of partnership
with machines.
Speaker 1 (34:26):
And finally wrapping up the showcase highlights Magic Lab. They
apparently put on a real powerhouse performance at WAIC, a
big statement.
Speaker 2 (34:36):
Oh absolutely if we connect this to the bigger picture.
Magic Lab delivered what many called a truly jaw dropping
live demo, just packed with speed, precision, next level coordination.
Speaker 1 (34:46):
What did their humanoids do?
Speaker 2 (34:48):
Impressive lifelike balance, flexibility. They showed them sprinting full out
then stopping instantly to catch something thrown at the mid stride.
Incredible predictive control and navigating crowds, definitely weaving through sem
related crowds, predicting human movements, avoiding collisions by centimeters, really
advanced social navigation.
Speaker 1 (35:06):
Robot dogs they.
Speaker 2 (35:07):
Took it even further. Synchronized tricks, backflips, navigating tough terrain
like something out of an action movie. Just amazing agility
and robust movement.
Speaker 1 (35:17):
So what was the key takeaway? Why was it such
a statement?
Speaker 2 (35:20):
What wowed people wasn't just the individual feats, but how
seamlessly they adapted to the dynamic environment. Weaving through crowds,
autonomously responding instantly to voice commands, working together in coordinated ways.
Speaker 1 (35:33):
It showed real world readiness.
Speaker 2 (35:35):
Exactly, a tangible leap forward in autonomy and readiness. Magic
Lab proved they're not just building robots for labs or
edited videos. They're building machines ready to perform reliably in public,
unpredictable spaces.
Speaker 1 (35:48):
So the message was the future is here.
Speaker 2 (35:51):
That was the feeling fast, flexible, autonomous, and surprisingly fun
and awe inspiring to watch, a clear signal that advanced
robots are rapidly into grind into our world.
Speaker 1 (36:01):
Okay, so, after this incredible journey through the absolute cutting
edge of AI robotics, what does it all mean for
you our listener?
Speaker 2 (36:09):
Yeah, let's try and synthesize.
Speaker 1 (36:10):
We've taken this deep dive into a real whirlwind of breakthroughs.
From affordable humanoids like Airli, opening up research and democratizing access.
Speaker 2 (36:19):
Making advanced robotics available to more innovators, to.
Speaker 1 (36:22):
These unified brains like skills promising universal intelligence across different.
Speaker 2 (36:27):
Robots, a huge potential accelerator.
Speaker 1 (36:29):
To multisensory perception systems like eximoids, letting machines truly understand
their environment in a much more humanlike.
Speaker 2 (36:37):
Way, moving beyond just a vision.
Speaker 1 (36:39):
And we've seen robots stepping directly into our world, serving
US drinks, doing tough, dangerous industrial jobs with precision, navigating
treacherous terrain, and even becoming well Maestro's on stage.
Speaker 2 (36:51):
Luring lines between utility and artistry.
Speaker 1 (36:53):
The common thread, the narrative weaving through all this seems
to be this relentless, accelerating push for autonomy, for adept ability,
and for real world integration.
Speaker 2 (37:02):
Getting robots out of the lab and into our lives.
Speaker 1 (37:04):
These aren't just experiments anymore. They're rapidly becoming practical tools,
efficient assistants, maybe even surprising collaborators, and they're set to
reshape industries, homes, and our basic understanding of what machines
can do and what role they'll play.
Speaker 2 (37:19):
And this incredible pace, it really does raise a vital,
compelling question for all of us looking ahead, doesn't it?
Speaker 1 (37:25):
Yeah, what's the big takeaway thought?
Speaker 2 (37:26):
As robots keep evolving like this, taking on roles we
once thought were impossible, for machines, becoming more capable, more ubiquitous,
we have to seriously consider what completely new opportunities and
let's be honest, what new challenges will emerge when these
intelligent companions become truly integrated into our society.
Speaker 1 (37:46):
How does it change things fundamentally?
Speaker 2 (37:48):
How will our understanding of collaboration, of efficiency, even of creativity?
How will all that continue to be redefined maybe expanded
by this deepening partnership reforming with these incredible machines.
Speaker 1 (38:00):
It's happening now.
Speaker 2 (38:01):
It's a future that's really not science fiction anymore. It's
becoming science fact, unfolding right now, right before our eyes,
and it kind of invites all of us to watch
closely and be part of this astonishing evolution.
Speaker 1 (38:15):
Well, we hope this deep dive has given you that
powerful shortcut to being well informed, maybe spark some aha
moments for you, or even some new questions about the
incredible possibilities ahead.
Speaker 2 (38:26):
There's certainly a lot to think about.
Speaker 1 (38:27):
The world of robotics is moving incredibly fast, and we'll
definitely be here tracking every mind blowing moment, every breakthrough.
So keep exploring, keep questioning, and definitely keep an eye
on what's coming next
Speaker 2 (38:39):
Because the future is arriving faster than you think.
Imagine stepping into a world where the future isn't just
you know, something on the horizon, it's already here, like
unfolding right now, right in front of us. Uh huh.
Picture advanced machines seamlessly serving you ice cream at an expo,
or effortlessly juggling your groceries right and even incredibly conducting
(00:21):
a live symphony orchestra with actual nuance and grace.
Speaker 2 (00:25):
It sounds like science fiction, doesn't it totally?
Speaker 1 (00:28):
And what if the most thrilling feats of say, athleticism,
weren't just about human skill anymore, but about the astonishing
agility of these intelligent machines.
Speaker 2 (00:39):
It's a huge shift.
Speaker 1 (00:40):
We're i mean, really living through a truly transformative time,
a moment where that line between what humans can do
and what machine intelligence can do is blurring faster than well,
faster than most of us can probably wrap our heads around.
Speaker 2 (00:52):
Absolutely, and it makes you wonder.
Speaker 1 (00:54):
Exactly, how are these machines doing such incredible things, things
we couldn't have imagined just a few years ago. What
are the real brain through is driving this just breathtaking pace?
Speaker 2 (01:03):
That's the core question.
Speaker 1 (01:04):
So today we're doing a deep dive We're looking at
some truly groundbreaking reports, some new insights that really show
the incredible speed of innovation in AI robotics, and this
deep dive it's designed for you, our listener, to get
through all the noise, pull out the most important, maybe
surprising bits.
Speaker 2 (01:25):
Of knowledge, and give it some context.
Speaker 1 (01:27):
Exactly with engaging context. Think of it like your shortcut
to being genuinely well informed on a topic that is,
let's face it, reshaping everything.
Speaker 2 (01:36):
It really is.
Speaker 1 (01:37):
Our mission here is pretty broad. We'll explore everything from
nimble humanoids doing daily tasks with surprising finesse.
Speaker 2 (01:44):
Yeah, the dexterity is getting amazing.
Speaker 1 (01:46):
Two powerful rugged robot dogs tackling complex, even dangerous terrains,
and crucially, the brain's behind it.
Speaker 2 (01:54):
All, the AI system's powering them. That's key.
Speaker 1 (01:57):
We'll unpack how robots aren't just evolving step by step
a truly revolutionizing things around us.
Speaker 2 (02:01):
It feels like an exponential curve.
Speaker 1 (02:03):
Sometimes right from making advanced robotics way more accessible, more
affordable than ever.
Speaker 2 (02:07):
Which is huge for innovation.
Speaker 1 (02:09):
To developing these unified brains that promise this universal intelligence
powering any robot for any task. That sounds wild, it's
a massive ambition and then showcasing just stunning real world
capabilities that honestly, they're going to make your jaw drop.
They really challenge your idea of what robots could even do.
Speaker 2 (02:30):
And what's truly fascinating here isn't just what they can
do now, but the how you know, the underlying tech. Yeah,
the mechanics, the software, yeah, exactly, the sophisticated control systems,
the multimodal perception letting them understand the world in a
much richer way, not just see it.
Speaker 1 (02:48):
Okay, multimodal perception. We'll get into that.
Speaker 2 (02:50):
And also the strategic shifts. Companies are making big, sometimes
audacious moves to get these machines out of the labs
and you know, into our everyday.
Speaker 1 (02:58):
Lives, right out into the real world.
Speaker 2 (03:00):
And this rapid progress. It brings up a really important
question for all of us, doesn't it. What does all
this mean for how we'll live, how will work, how
we'll interact with technology in the well, the very near future.
Speaker 1 (03:10):
It's not abstract anymore.
Speaker 2 (03:12):
Not at all. It's a question we need to be
thinking about because these intelligent machines are moving fast from
theory and lab tests to actual tangible things impacting our
lives daily.
Speaker 1 (03:24):
Okay, let's dive into the first big piece of news, then,
the one that's got the robotics world buzzing. That's the
official launch of Limix Dynamics new humanoid robot Oli ah.
Speaker 2 (03:34):
Yes, Oli formerly CL three that used to.
Speaker 1 (03:37):
Be CL three, And it's making waves not just for
what it can do, but maybe even more for its
price tag, which kind of stunning really is. We're taking
a starting price around one hundred and fifty eight thousand,
which translates to roughly twenty two thousand US dollars.
Speaker 2 (03:52):
Twenty two thousand for a humanoid.
Speaker 1 (03:54):
That's not just a number, right, that feels like a statement. Yeah,
So for those of us watching this field accelerate, what
is that price point actually mean for the market for accessibility?
Speaker 2 (04:03):
Well, if you connect this to the bigger picture, it's
genuinely significant. I'd even say transformative for the industry.
Speaker 1 (04:08):
Ysorm of hew think about it.
Speaker 2 (04:10):
For a fully capable humanoid robot. It stands five foot
five styles so human.
Speaker 1 (04:14):
Size basically fits in human environments.
Speaker 2 (04:17):
Exactly, weighs fifty five kilos, which is manageable, and it
has thirty one degrees of freedom.
Speaker 1 (04:22):
Okay, thirty one degrees of freedom? Break that down quickly.
What does that allow?
Speaker 2 (04:25):
It means an extraordinary range of motion. Think about your arm,
maybe seven degrees of freedom. Lets you do all sorts
of things, thirty one across the whole body. That enables
incredibly complex, nuanced, almost human like movements, fine motor control,
navigating cluttered spaces gracefully.
Speaker 1 (04:42):
Got it so, highly capable.
Speaker 2 (04:45):
Highly capable, and at that price point twenty two thousand dollars,
that positions o ALI as a truly affordable option.
Speaker 1 (04:52):
Affordable compared to what typically what were we looking at?
Speaker 2 (04:55):
Multiples of that hundreds of thousands easily for research grade
humanoids with similar specs, often millions for the really high
end ones.
Speaker 1 (05:02):
Wow.
Speaker 2 (05:03):
Okay, So this affordability is the key. It means OLILI
is accessible not just to the big tech giants with
massive R and D budget.
Speaker 1 (05:10):
Right at the usual suspect.
Speaker 2 (05:12):
But crucially to researchers, innovative startups, universities all over the world.
It effectively democratizes access developing on advanced humanoid platforms.
Speaker 1 (05:20):
So it spreads the innovation potential around precisely.
Speaker 2 (05:23):
We're seeing these incredible machines move beyond being exclusive toys
for a few big players. It opens them up to
a much broader base of innovators who are you know,
really eager to push the boundaries now that they can
actually get their hands on the hardware.
Speaker 1 (05:37):
That is genuinely remarkable. But when you hear fully capable
at such a low price twenty two k and does
make you wonder how do they manage that? What are
the technical choices that let them hit that price without
you know, sacrificing core abilities.
Speaker 2 (05:53):
That's a great question, And one key thing attestament to
some clever engineering for cost efficiency is OBII in house
developed six axis.
Speaker 1 (06:02):
IMU im you inertial measurement units exactly.
Speaker 2 (06:05):
Now. IMU sounds super technical, I know, but it's absolutely crucial.
It's like the robot's inner ear, especially for a bipedal
one that needs to balance.
Speaker 1 (06:13):
Like how we stay upright without thinking about.
Speaker 2 (06:15):
It pretty much, imagine trying to walk on two legs
if you didn't constantly know how your body is tilted,
how fast it's turning or accelerating. The IMU provides that
precise real time data orientation, angular velocity, linear acceleration.
Speaker 1 (06:28):
All constantly feeding into the control system constantly.
Speaker 2 (06:31):
It's vital for precise motion control and most importantly for
dynamic balance. For all at LIE to walk smoothly stably,
especially in human spaces, it needs to constantly monitor and
instantly adjust.
Speaker 1 (06:44):
So developing that sophisticated piece in house was probably a
big cost saver while keeping performance high.
Speaker 2 (06:51):
Very likely, it gives them direct control over a core component,
ensuring it meets their needs for precise movement, without relying
on expensive off the shelf parts that might have features
they don't need or not integrate as well.
Speaker 1 (07:04):
Okay, makes sense. And what about for the people who
want to actually use all I, the developers, the researchers,
the students. How easy is it for them to get started?
Speaker 2 (07:12):
This is where Limus Dynamics really seems to be playing
it smart, showing a very forward thinking strategy. ALLI comes
with a modular.
Speaker 1 (07:19):
SDK software development kit right.
Speaker 2 (07:20):
And a fully open interface, and it allows for Python
based customization.
Speaker 1 (07:24):
And control Python. Okay, that's huge.
Speaker 2 (07:26):
It's a game changer for the developer community. It means
they aren't locked into some proprietary complex system. They can
program OLI using Python, which is one of the most popular,
versatile developer friendly languages out there, so.
Speaker 1 (07:40):
The learning curve is lower, the potential developer pool is massive.
Speaker 2 (07:44):
Exactly. It supports rapid iteration. Developers can quickly test new ideas,
experiment with different applications, integrate OLI into diverse scenarios from
fundamental AI research to practical service robotics.
Speaker 1 (07:57):
So limics isn't just selling a robot. They're selling a platform.
Speaker 2 (08:00):
That seems to be the strategy, a development platform, almost
like an advanced computer you can program for tons of
different tasks, not just a fixed function machine. This openness
plus the price, plus the real time control, it makes
OLI a serious disruptor, especially.
Speaker 1 (08:16):
For anyone looking for that balance between high capability and
genuine accessibility. Right.
Speaker 2 (08:21):
It kind of invites you, the listener to think, Okay,
if I had access to this, what could I build?
Speaker 1 (08:25):
It?
Speaker 2 (08:25):
Fosters innomation way beyond Limits's own labs.
Speaker 1 (08:28):
Okay, so OILI is out there making waves with affordability.
But then there's the anticipation game. Fourier Intelligence just dropped
a sneak peak of its upcoming GR three humanoid.
Speaker 2 (08:39):
Ah the GO three teaser. Yeah, that got people talking.
Speaker 1 (08:42):
It's just a digital render, not the real thing yet.
But enough to create buzz ahead of the full reveal
on August sixth. What's the strategy here? Why release just
a render?
Speaker 2 (08:53):
Well, what's fascinating is the timing and the subtle hints
in that render. They did show a prototype briefly at
WAIC twenty twenty five in Shanghai.
Speaker 1 (09:03):
The event will definitely circle back to for.
Speaker 2 (09:04):
Sure, But this new render, it's suggest something sleeker, more
advanced and evolution, moving maybe from just function towards a
more refined design.
Speaker 1 (09:13):
So what hints are we seeing? What is refined suggest the.
Speaker 2 (09:15):
Focus seems to be on well, refine mobility, smoother, more
precise movements, better agility, also enhance mechanical balance, which implies
more stability, especially when things get unpredictable, and potentially more
sophisticated AI interaction capabilities.
Speaker 1 (09:29):
Meaning it could understand and respond.
Speaker 2 (09:31):
Better, possibly respond more intelligently, more intuitively to its environment,
maybe to human commands. We don't know for sure yet,
but that's the vibe.
Speaker 1 (09:40):
And releasing this teaser just two weeks before launch pretty
standard hype building.
Speaker 2 (09:46):
It's smart marketing in a very competitive field. It puts
Fourier intelligence right back in the conversation in this rapidly
accelerating humanoid race. You know, companies in China, globally, everyone's
pushing hard for real world cap robots at scale.
Speaker 1 (10:01):
So it's about signaling leadership, attracting attention.
Speaker 2 (10:04):
Exactly, building anticipation, signaling they're a major player, attracting potential partners,
creating a narrative. If this renders anything to go by,
the g R three could be a pretty significant step
for Fouria, maybe their most important humanoid yet.
Speaker 1 (10:18):
Okay, so from anticipation to something we've actually seen in action,
let's talk about Droidip's Walker Too. It was already kind
of a fan favorite, right, this bipedal roma.
Speaker 2 (10:26):
Yeah, Walker two has definitely had a following.
Speaker 1 (10:28):
It just got a major upgrade, Yeah, autonomous navigation. But
the surprise was they also introduced a new companion, Rina,
the wheeled one, right, and together they put on this, well,
this synchronized show. Beyond just looking cool, what does that
performance tell us about where multi robot collaboration is heading.
Speaker 2 (10:48):
This brings up a really interesting point about how robots
are moving beyond just doing their own thing towards expressive
and really intricate coordination like a team. Exactly, Walker Too
can how navigate dynamic places all by itself. It responds
in real time, keeps its balance, understands the space around it.
It's not just following lines on the.
Speaker 1 (11:08):
Floor, it's perceiving and deciding right.
Speaker 2 (11:11):
And Rena the wheeled one. She's built for speed and
smooth movement on flat surfaces, very agile, so different strengths,
totally different strengths, and their duet style dance routines at
WWAC twenty twenty five weren't just for show. They really
showcased Droydep's mastery in getting different types of robots to collaborate.
Speaker 1 (11:27):
Two different locomotion systems working together precisely.
Speaker 2 (11:30):
Imagine that a bipedal robot good for human spaces and
a wheeled one good for speed, performing a complex synchronized routine,
reacting to each other, reacting to the music, the environment seamlessly.
Speaker 1 (11:43):
That coordination must be incredibly complex to program.
Speaker 2 (11:46):
Oh. Absolutely. The contrast between Walker two sort of bipedal
elegance and Rena's smooth rolling agility really highlighted their synergy.
It showed how different robot types can complement each other.
Speaker 1 (11:58):
And the implications go beyond dance through teams. Obviously, for sure.
Speaker 2 (12:01):
Think about complex operations in the future, logistics and a
busy warehouse, maybe disaster response. It probably won't be one
single robot doing everything.
Speaker 1 (12:10):
It'll be teams, heterogeneous teams.
Speaker 2 (12:13):
Exactly specialized machines working together, some agile, some strong, some fast,
optimizing the whole system. This demo hnts that future where
robots coordinate complex tasks in industry, maybe in service roles,
or even become more engaging companions at home.
Speaker 1 (12:28):
Okay, shifting gears a bit. One clip that really went
viral from WAIC twenty twenty five involved Unitre's g one
humanoid ah.
Speaker 2 (12:37):
The fighting robot demo. Yeah, that blew up online.
Speaker 1 (12:39):
It sounds like pure sci fi. The g One squaring
off against professional fighters. You'd think it's a mismatch, but
apparently the g one held its own.
Speaker 2 (12:47):
Held its own is putting it mildly. What was fascinating
was the sheer, physical prowess, the agility the g One displayed.
It has twenty six high performance.
Speaker 1 (12:56):
Joints, which is a lot of control points, a lot.
Speaker 2 (12:59):
And it demonstrated this martial arts style agility that honestly
stunned people in the videos you see it blocking ducking,
striking back with shockingly fluid motions like it knew it
was coming, almost as if it was reading the opponent's
moves in real time. And these were professional fighters, not actors.
They were going at it, but the g One kept
its robotic composure, didn't fall over, didn't.
Speaker 1 (13:21):
Break, so definitely more than just a fun stunt.
Speaker 2 (13:23):
Way more, it highlighted how far humanoid robotics has come
in motion control, in reflex training. The g One wasn't
just walking or doing preset moves. It was sparring with
trained athletes, dynamic interaction, physical resilience. Stuff that was pure science.
Speaker 1 (13:38):
Fiction very recently, and it sparked debates right.
Speaker 2 (13:40):
Oh yeah, across tech circles, even in combat sports. People
were asking, are we seeing the dawn of robotic fighters
or is this mainly just a very very impressive demo
of motion control and reaction speed.
Speaker 1 (13:51):
Either way, it shows a huge leap in capability.
Speaker 2 (13:54):
A full blown tech spectacle, absolutely yeah, proving robots can
do these high speed hidexs starity things with surprising control
and toughness.
Speaker 1 (14:03):
When you say shocking fluidity and reading moves, that sounds
like incredibly advanced AI and control behind the scenes. What
kind of tech makes that possible?
Speaker 2 (14:13):
Yeah, to get that kind of fluidity and real time reaction,
you're looking at extremely advanced, low latency control systems, almost
certainly using sophisticated AI, maybe real time reinforcement learning or
predictive modeling.
Speaker 1 (14:27):
So it's not just reacting, it's anticipating.
Speaker 2 (14:29):
Potentially, it needs incredibly fast sensor feedback, high speed cameras,
probably force and torque sensors in the joints and limbs.
The AI has to process all that data instantly to
generate the right countermove. It's not just speed, it's intelligent
adaptive response.
Speaker 1 (14:44):
And the fact that it didn't you fall apart, right.
Speaker 2 (14:46):
That implies a really robust mechanical design, too able to
take impacts and efficient power management to keep going during
that intense activity.
Speaker 1 (14:54):
So beyond the wow factor of fighting robots, where else
could this kind of agility be used? What other high
dexterity applications could this unlock?
Speaker 2 (15:03):
Oh? Tons of places? Imagine robots with this agility doing
complex assembly and high tech manufacturing, placing tiny components with
split second precision.
Speaker 1 (15:13):
Faster and maybe more accurate than.
Speaker 2 (15:15):
Humans potentially, or think about emergency response robots navigating rubble,
unstable buildings, clearing debris, maybe even performing rescues in situations
too dangerous.
Speaker 1 (15:26):
For people, hazardous environments too, like nuclear sites or chemical handling.
Speaker 2 (15:30):
Exactly performing intricate tasks with precision, endurance and safety that
humans just can't match. This level of agility, precision, resilience.
It could fundamentally change fields needing fine motor skills, fast reactions,
and tough physical performance, logistics, healthcare construction. The list goes on.
Speaker 1 (15:49):
Okay, so we've covered some amazing physical feats, agility, coordination, resilience,
but the body is only half the story, maybe less.
The real revolution is often in the brain.
Speaker 2 (15:59):
The intelligence driving it all.
Speaker 1 (16:00):
Absolutely, So let's shift our deep dive there to the intelligence,
the perception, the big ambitions to create unified AI for
all robots. After staying quiet for months, Skilled finally gave
a glimpse into their huge project.
Speaker 2 (16:16):
Yeah. Skilled emerged from Stealth last year, and now we're
hearing more.
Speaker 1 (16:20):
They're trying to build a single general purpose robotics foundation
model one AI to power any robot for any task.
Speaker 2 (16:29):
That sounds monumental, almost sci fi. Why does this Skilled brain.
Speaker 1 (16:34):
Well connecting it to the bigger picture. It's a truly
audacious vision, potentially revolutionary for robotics. The skilled brain is
basically a unified control system designed to operate different kinds of.
Speaker 2 (16:45):
Robots, humanoids, quadrupeds.
Speaker 1 (16:47):
Arms exactly humanoids likeli those quadrupeds we talked about, even
simpler robotic arms, all using the same underlying AI architecture.
Speaker 2 (16:55):
How is that even possible? Usually they need separate programming.
Speaker 1 (16:57):
Right, that's a traditional way, but skilled different think about
how we learn. We learn general principles about picking things up, physics,
object properties, how to grip, and we apply that whether
we're picking up a feather or a brick with our
hands or a tool. So they want robots to have
that same kind of general understanding.
Speaker 2 (17:14):
That's the goal, generalized intelligence for robots, and they're doing
it by training their model on massive, incredibly diverse data.
Speaker 1 (17:22):
Sets of actions, well kind of data.
Speaker 2 (17:24):
Everything, huge amounts of Internet videos showing humans and machines
doing things in all sorts of contexts, high fidelity simulation
data where they can learn safely and quickly in virtual worlds,
and tons of real world sensor data from actual robots
doing actual tasks.
Speaker 1 (17:40):
So quantity and variety of data is key.
Speaker 2 (17:42):
Absolutely crucial. The goal is to break down those traditional
silos in robot control where you train one model for walking,
another for grasping, another for a.
Speaker 1 (17:50):
Different robot, which is slow and expensive.
Speaker 2 (17:53):
Incredibly slow and expensive.
Speaker 1 (17:54):
Yeah.
Speaker 2 (17:55):
Skilled wants a scalable, adaptable intelligence that generalizes across different
robot bodies, a single brain that could theoretically run a
warehouse robot one minute and a home assistant the next.
Speaker 1 (18:05):
Which raises the question could this really speed up robot
development everywhere?
Speaker 2 (18:10):
That's the billion dollar question, isn't it. If they succeed,
it could massively accelerate things, making it faster and cheaper
to deploy smart robots into completely new areas. They're aiming
to build the universal intelligence to run them all. It
could fundamentally reshape the entire.
Speaker 1 (18:26):
Field that lead from bespoke AI for each robot to
a universal model. Why is breaking silos so critical for
robotics and what makes Skill think they can pull off
such a huge challenge.
Speaker 2 (18:38):
Well, the traditional way, being so fragmented, is a huge
bottleneck developing separate models for every robot every task. It's
resource intensive, slow, hard to scale. Imagine needing a unique
brain for every single thing you want a robot.
Speaker 1 (18:52):
To do, progress would crawl exactly.
Speaker 2 (18:55):
A unified foundation model, though, learns general principles of interacting
with the physical world from watching countless examples. It's these
patterns generalizes skills, kind of like how large language models
llms revolutionized language.
Speaker 1 (19:07):
Tasks like GBT for language right.
Speaker 2 (19:09):
Instead of training separate models for translation, summarization Q and
A one big foundational LLM can adapt skilled. Is trying
to do the same for physical actions, building essentially a
GPT for robots learning the grammar and physics of interaction
from massive.
Speaker 1 (19:25):
Data so the data is absolutely fundamental.
Speaker 2 (19:28):
Data is king is an understatement here. Training on those massive,
diverse action data sets is how they give the AI
a comprehensive, nuanced understanding of how things work. Objects forces environments.
Speaker 1 (19:40):
Like learning to open any door, not just one specific.
Speaker 2 (19:43):
Door, precisely seeing thousands of examples different doors handles people
opening them. The physics involved that breadth and variety are
crucial for robots to be truly robust and adaptable and messy,
unpredictable real world situations and skilled itself.
Speaker 1 (19:57):
What's their background?
Speaker 2 (19:58):
What makes them compelling is the team founded in twenty
twenty three by really distinguished professors from Carnegie Mellon.
Speaker 1 (20:04):
Universities CMU, a powerhouse in robotics.
Speaker 2 (20:07):
Research, absolutely and backed by robotics pioneers who've done some
of the most important work in the field over the
last decade. This isn't just a startup with a flashy idea.
It's deep tech with a very strong academic and research foundation.
Speaker 1 (20:21):
And now they're shifting from research to application.
Speaker 2 (20:24):
Exactly taking that deep knowledge and applying it to this
massive real world problem. They're not just building robots, they're
aiming to build the universal brain for robots, the foundational
intelligence that could unlock a whole new generation of capabilities.
Speaker 1 (20:40):
Okay, so that's the brain. Now, how do robots understand
the world To feed that brain?
Speaker 2 (20:46):
We usually think cameras, right, vision visions is definitely dominant historically.
Speaker 1 (20:50):
Meximoyd just unveiled something called humanoid Occupancy. They're calling it
the world's first general purpose multimodal perception system specifically for humanoids.
Speaker 2 (21:00):
That sounds like more than just cameras.
Speaker 1 (21:02):
It is much more, and what's fascinating is how it
tries to mirror or even extend human perception you're right.
Traditional systems rely heavily on cameras.
Speaker 2 (21:10):
Robots mostly see, but we use more than just our eyes,
way more we use site sure, but also hearing, detecting sounds,
figuring out where they're coming from, what they mean, touch, feeling, textures, forces, temperature,
and our spatial awareness appropriateception knowing where our body parts
are without looking.
Speaker 1 (21:29):
Okay, so humanoid occupancy tries to give robots all of that.
Speaker 2 (21:33):
That's the idea, integrating audio, tactile, touch, and spatial sensing
into one unified system. And it's not just about adding
more sensors. The key is sensor.
Speaker 1 (21:43):
Fusion, combining the data intelligently.
Speaker 2 (21:45):
Intelligently combining and interpreting data from all these different sources visions, sound, touch,
spatial awareness in real time to build a truly holistic,
robust internal picture of the environment.
Speaker 1 (21:57):
Can you give an example, how does that help?
Speaker 2 (21:58):
Sure? Imagine a row in a busy factory. Its camera
view might be blocked by a stack of boxes for
a second.
Speaker 1 (22:04):
A vision only robot would be blind, right.
Speaker 2 (22:07):
But a multimodal robot might hear the distinct sound of
an approaching forklift. Its tactile sensors might feel a slight
vibration in the floor as the heavy vehicle gets closer.
Speaker 1 (22:17):
So even without seeing it, it knows something's coming exactly.
Speaker 2 (22:20):
By fusing these different inputs, it gets a much richer,
more human like understanding. It can make safer, more nuanced
decisions than a robot just relying on site.
Speaker 1 (22:30):
This sounds like the foundation for well more intelligent behavior.
Speaker 2 (22:34):
It really is. It lays the groundwork for a new
era of embodied intelligence, robots moving from just reacting to
what they see to making context to were decisions precisely
based on a fuller understanding of the world, whether they're
in homes, factories, outlowers. It's a huge shift from reactive
automation just doing pre programmed things or reacting to visual cues,
(22:55):
to proactive intelligent assistance.
Speaker 1 (22:57):
Where they can anticipate, adapt make better.
Speaker 2 (23:00):
Judgment yes, which dramatically improves their safety, their efficiency, their adaptability,
especially in unpredictable real world settings. It's a profound step up.
Speaker 1 (23:09):
All these individual breakthroughs, affordable hardware, unified brains, multimodal perception,
they really seemed to come together at one big event
WAIC twenty twenty five in Shane High.
Speaker 2 (23:20):
Oh Yeah, WIC was huge this year, a real landmark.
Speaker 1 (23:23):
Described as the largest embodied AI and robotics showcase ever.
It painted this vivid picture of the future actually happening.
What really stood out there? How did it show the
convergence of everything we've been talking about.
Speaker 2 (23:36):
WAIC twenty twenty five was honestly a powerful statement. Embodied
AI isn't niche anymore. It's taking center stage globally. The
scale was just breathtaking.
Speaker 1 (23:47):
Give us some numbers.
Speaker 2 (23:48):
Over three thousand AI innovations showcased, more than one hundred
and fifty different humanoid robots and over sixty brand new
robot models debuting all in one place. It felt like
a turning point.
Speaker 1 (24:00):
And the robots themselves what was different.
Speaker 2 (24:02):
They weren't just slightly more capable. They were noticeably smarter, faster,
and interestingly more socially aware.
Speaker 1 (24:08):
Socially aware how so.
Speaker 2 (24:09):
Meaning they could interact with humans and navigate crowded spaces
with more nuanced, more understanding. They showed greater autonomy, adapted
better to new situations, handled complex environments more easily. Interactions
felt more fluid, more intuitive.
Speaker 1 (24:22):
So the theme was really about readiness for the real world.
Speaker 2 (24:25):
Absolutely, the overarching themes were autonomy, adaptability, and real world integration.
This wasn't just a show of lab prototypes that might
be useful someday. It felt like a showcase of machines
ready or very nearly ready to operate among us in
our daily lives.
Speaker 1 (24:42):
And this cannakes back to things like skilled unified brain
and Eximoid's multimodal perception.
Speaker 2 (24:47):
Directly, those concepts a universal brain, rich multisensory understanding aren't
just features. There are critical enablers for the kinds of autonomous, adaptable,
socially aware robots we saw everywhere at WAIC.
Speaker 1 (25:00):
WAIC proved the pieces are coming together.
Speaker 2 (25:02):
It vividly demonstrated that the next wave is about machines
truly being integrated into our environments, doing diverse tasks intelligently, fluidly, independently.
The sheer scale and variety of innovation it signaled a
field that's mature and accelerating incredibly fast. We're on the
cusp of a major robotic revolution impacting well everything.
Speaker 1 (25:21):
Let's zoom in on that real world impact. At WAIC
we saw robot stepping directly into service and retail rolls,
moving beyond factories. Tell us about the Pepsi cool bot
and that robotic mini store concept.
Speaker 2 (25:32):
Right, those were fascinating because they showed immediate, tangible practicality
robots moving into public consumer facing jobs.
Speaker 1 (25:40):
So the Pepsi cool bot, what was that.
Speaker 2 (25:42):
It was a collaboration between a Gibbet and Pepsi. And
this robot wasn't just static. It autonomously navigated the busy
expo floor, weaving through crowds, yeah, with impressive precision. It
even had a certain charm in its movements, apparently delivering
drinks directly to people.
Speaker 1 (25:58):
So not just a gimmick, it didn't.
Speaker 2 (25:59):
Feel like one, and it felt like a clear glimpse
into the future of retail automation, handling, distribution, inventory, even
customer interaction in a dynamic public space.
Speaker 1 (26:09):
And the robotic mini store that.
Speaker 2 (26:11):
Was from Gelbit Robotics basically a fully autonomous retail kiosk,
a self contained little store. Visitors could walk up, interact
with it, browse products, choose something, and complete the purchase,
all without any human staff.
Speaker 1 (26:25):
Wow, like a really advanced vending machine.
Speaker 2 (26:28):
Kind of, but much more interactive and flexible. Again, these
aren't just concepts. There are operational previews of a more
automated efficient, maybe more convenient retail landscape.
Speaker 1 (26:38):
What are the implications here for businesses and consumers?
Speaker 2 (26:41):
For businesses, potential for optimized operations, lower staffing costs maybe
two hundred and four to seven. Service For consumers maybe
instant frictionless shopping, more access to goods. It points to
a future where convenience and efficiency are increasingly driven by
smart automation. It changes how business is run and how
we shop.
Speaker 1 (27:02):
Okay, Moving from retail to the tougher environments factories, warehouses,
WAIC showed big strides there too. Leisure Robotics had new
materials sorting tech.
Speaker 2 (27:13):
Yes, Lijiu showed a major upgrade. Their humanoids can now recognize, classify,
and organize a wide variety of materials in.
Speaker 1 (27:21):
Real time, beyond just color and shape.
Speaker 2 (27:23):
Way beyond. In the demos, the robot was accurately distinguishing
between plastics, different metals, fabrics, glass, sorting them into bins flawlessly.
Speaker 1 (27:30):
Does it do that?
Speaker 2 (27:31):
It uses advanced visual recognition, but also AI driven texture analysis,
maybe other sensors, and contextual reasoning to figure out what
it's holding.
Speaker 1 (27:39):
And why is that important? Material sorting?
Speaker 2 (27:41):
It's a hugely tedious repetitive, often dirty, and sometimes hazardous
job for humans, especially in recycling plants or waste management.
Lijiu's tech shows AI turning humanoids from show pieces into
practical helpers for these tough tasks, reduces human burden boost
sufficiency could transform large scale res cycling, maybe even home
waste separation down the line.
Speaker 1 (28:03):
Then there was AI, two robotics and their humanoid Aibo
doing palatizing.
Speaker 2 (28:09):
Aibo showcased whole domain full body palatizing, and it was
seriously impressive.
Speaker 1 (28:14):
What does full body palatizing mean?
Speaker 2 (28:16):
It means the robot used its entire body torso rotation,
core strength, fine finger control to lift and stack boxes
onto a palette super efficiently in under sixty seconds. It
stacked multiple boxes perfectly.
Speaker 1 (28:28):
How did it manage that coordination?
Speaker 2 (28:30):
It dynamically adjusted its posture, anticipated how the boxes would shift,
fine tuned its grip in real time, all autonomously using
its visual sensors and AI coordination. No resets, no hesitation,
no dropped boxes, just flawless execution.
Speaker 1 (28:43):
That sounds like a game changer for logistics.
Speaker 2 (28:45):
Absolutely, palatizing is physically demanding, high risk for human injury.
Aibo showed robots can do it faster, more consistently, maybe safer.
It's a clear preview of the next gen factory or
distribution center and.
Speaker 1 (28:58):
Chinten Cyborg roboticslaunched their ro one humanoid.
Speaker 2 (29:01):
Yes, the Cyborg R one made its official debut. Sleek
industrial design but clearly built for real world function.
Speaker 1 (29:09):
What could it too?
Speaker 2 (29:09):
It showed smooth walking, stable balancing, precise hand control, using tools,
transferring objects, even responding to gestures naturally.
Speaker 1 (29:18):
What makes it adaptable?
Speaker 2 (29:19):
It has a modular architecture and advanced AI high performance joints,
good three D spatial awareness and natural language processing so
it can understand complex voice commands.
Speaker 1 (29:28):
So it could be used for lots of different factory
or service jobs.
Speaker 2 (29:31):
That's the idea multipurpose assistant. Its plug and play flexibility
makes it adaptable for different tasks. This whole suite of
industrial robots from le Zeu AI two Cyborg it signals
a real acceleration towards useful task specific AI turning humanoids
into indispensable real world workers.
Speaker 1 (29:49):
Humanoids get a lot of attention, but the robot dogs
the quadrupeds are also making huge leaps right, especially in tough.
Speaker 2 (29:55):
Spots definitely quadruped progress is incredibly impressive and arguably more
practical in many near term scenarios.
Speaker 1 (30:02):
We saw Rambo from ETH Zurich doing multitasking.
Speaker 2 (30:07):
Yeah. Rambo was amazing, developed by gin change In's team
at ETH. It's stunned people by pushing a shopping cart
while balancing a plate.
Speaker 1 (30:13):
Brunch hustling robot Boss.
Speaker 2 (30:15):
Exactly look charming effortless, but underneath is incredibly sophisticated control.
Rambo stands for RL augmented model based whole body control
for loco manipulation.
Speaker 1 (30:27):
Okay, unpack that acronym RL augmented.
Speaker 2 (30:30):
It uses a hybrid approach precise model based planning for
accuracy combined with real time reinforcement learning feedback.
Speaker 1 (30:36):
So it learns and adapts on the fly right.
Speaker 2 (30:38):
Which makes it super robust and adaptable in dynamic situations.
It lets Rambo seamlessly combine walking and manipulation, which is
a huge challenge.
Speaker 1 (30:46):
In robotics, handling complex coordination naturally.
Speaker 2 (30:49):
Yeah, adjusting to uneven floors, recovering from bumps, balancing things
while moving. It showed amazing fluidity and resilience, a big
leap in mobile manipulation and.
Speaker 1 (31:00):
Deep Robotics showed a whole range of quadrupeds.
Speaker 2 (31:03):
They did a real showstopper, from smaller ones for tight
spaces to heavy duty industrial models, impressive power agility, real
world functionality. So was their focus embodied AI. These robots
weren't just following commands, They were thinking, using perception, real
time decision making, adaptive learning, what kind of tax, navigating
(31:24):
complex obstacles, climbing stairs, easily carrying heavy payloads over rough ground,
responding to voice and gestures intuitively. They showed uses from
power inspections and emergency response to logistics and construction, proving
their AI platform is scalable and versatile.
Speaker 1 (31:39):
They seem ready for serious work.
Speaker 2 (31:41):
Absolutely Deep Robotics basically brought the future to the show floor,
proving quadrupeds with embodied AI are already solving real problems
with speed, accuracy, and autonomy.
Speaker 1 (31:50):
And then there was the heavy lifter phi In Titan
from the Southern Power Grid.
Speaker 2 (31:55):
Yes, py On Titan is breaking records designed specifically for
power grid inspection. It tackles the load bearing limits of
older robots.
Speaker 1 (32:04):
How does it manage heavy loads and tough terrain.
Speaker 2 (32:06):
It has an innovative hybrid mobility system wheels and legs.
Speaker 1 (32:10):
Best of both worlds, kind.
Speaker 2 (32:12):
Of unmatched adaptability across gravel slopes, industrial debris places critical
for inspections but often dangerous or inaccessible for humans or
wheeled robots.
Speaker 1 (32:22):
And it carries tools exactly.
Speaker 2 (32:24):
It's a platform for specialized gear HD, thermal cameras, lightar,
for three D mapping, robotic arms, other sensors. It excels
at fighting defects, reading meters, autonomous.
Speaker 1 (32:35):
Patrolling with AI helping identify problems.
Speaker 2 (32:37):
Yes, Its AI recognition systems spot abnormalities in real time,
reduces manual labor, boosts safety dramatically and high voltage areas.
It's a major leap for smart grid maintenance and autonomous
robotics and critical infrastructure, transforming how power inspections get done.
Speaker 1 (32:52):
Okay, from heavy industry to something completely different. Astrobots s
one robot performing with the Beijing National Orchestra as a
conductor and music.
Speaker 2 (33:00):
Yeah, that one truly blurred the lines automation meets artistry.
It sounds completely sci fi, doesn't it?
Speaker 1 (33:07):
A robot conductor? Totally unpredictable? What made it so jaw dropping?
What did it actually do?
Speaker 2 (33:12):
It raises fascinating questions about performance, collaboration, even emotion in
human robot interaction. The S one guided the orchestra with
remarkably accurate, fluid conducting.
Speaker 1 (33:23):
Gestures, setting tempo dynamics, queuing sections, all in real time.
Speaker 2 (33:27):
And it wasn't just rigid, pre programmed movements. It was dynamic,
responsive to the live musicians.
Speaker 1 (33:33):
It adapted to the orchestra.
Speaker 2 (33:34):
That's what was so impressive. It seemed to genuinely interact,
responding to subtle musical shifts. It even acknowledged the audience.
People felt a surprisingly emotional connection, a real sense of presence.
Speaker 1 (33:45):
Wow. Not just robotic repetition.
Speaker 2 (33:47):
Not at all. It was a profound proof of concept
showing humanoids can engage in high level collaboration, meeting rhythm timing,
nuanced human interaction, from smooth wrist flicks conveying musical ideas
to keeping the whole orchestra together. It blurred that line
between automation and artistry beautifully.
Speaker 1 (34:05):
So robots and cultural experiences.
Speaker 2 (34:07):
It marks a new chapter, robots not just working behind
the scenes, but becoming integral, maybe even inspiring parts of
culture and art, showing their learning to move, feel, perform
like us suggests the future of robots as genuine collaborators
in creative fields. It really expands your idea of partnership
with machines.
Speaker 1 (34:26):
And finally wrapping up the showcase highlights Magic Lab. They
apparently put on a real powerhouse performance at WAIC, a
big statement.
Speaker 2 (34:36):
Oh absolutely if we connect this to the bigger picture.
Magic Lab delivered what many called a truly jaw dropping
live demo, just packed with speed, precision, next level coordination.
Speaker 1 (34:46):
What did their humanoids do?
Speaker 2 (34:48):
Impressive lifelike balance, flexibility. They showed them sprinting full out
then stopping instantly to catch something thrown at the mid stride.
Incredible predictive control and navigating crowds, definitely weaving through sem
related crowds, predicting human movements, avoiding collisions by centimeters, really
advanced social navigation.
Speaker 1 (35:06):
Robot dogs they.
Speaker 2 (35:07):
Took it even further. Synchronized tricks, backflips, navigating tough terrain
like something out of an action movie. Just amazing agility
and robust movement.
Speaker 1 (35:17):
So what was the key takeaway? Why was it such
a statement?
Speaker 2 (35:20):
What wowed people wasn't just the individual feats, but how
seamlessly they adapted to the dynamic environment. Weaving through crowds,
autonomously responding instantly to voice commands, working together in coordinated ways.
Speaker 1 (35:33):
It showed real world readiness.
Speaker 2 (35:35):
Exactly, a tangible leap forward in autonomy and readiness. Magic
Lab proved they're not just building robots for labs or
edited videos. They're building machines ready to perform reliably in public,
unpredictable spaces.
Speaker 1 (35:48):
So the message was the future is here.
Speaker 2 (35:51):
That was the feeling fast, flexible, autonomous, and surprisingly fun
and awe inspiring to watch, a clear signal that advanced
robots are rapidly into grind into our world.
Speaker 1 (36:01):
Okay, so, after this incredible journey through the absolute cutting
edge of AI robotics, what does it all mean for
you our listener?
Speaker 2 (36:09):
Yeah, let's try and synthesize.
Speaker 1 (36:10):
We've taken this deep dive into a real whirlwind of breakthroughs.
From affordable humanoids like Airli, opening up research and democratizing access.
Speaker 2 (36:19):
Making advanced robotics available to more innovators, to.
Speaker 1 (36:22):
These unified brains like skills promising universal intelligence across different.
Speaker 2 (36:27):
Robots, a huge potential accelerator.
Speaker 1 (36:29):
To multisensory perception systems like eximoids, letting machines truly understand
their environment in a much more humanlike.
Speaker 2 (36:37):
Way, moving beyond just a vision.
Speaker 1 (36:39):
And we've seen robots stepping directly into our world, serving
US drinks, doing tough, dangerous industrial jobs with precision, navigating
treacherous terrain, and even becoming well Maestro's on stage.
Speaker 2 (36:51):
Luring lines between utility and artistry.
Speaker 1 (36:53):
The common thread, the narrative weaving through all this seems
to be this relentless, accelerating push for autonomy, for adept ability,
and for real world integration.
Speaker 2 (37:02):
Getting robots out of the lab and into our lives.
Speaker 1 (37:04):
These aren't just experiments anymore. They're rapidly becoming practical tools,
efficient assistants, maybe even surprising collaborators, and they're set to
reshape industries, homes, and our basic understanding of what machines
can do and what role they'll play.
Speaker 2 (37:19):
And this incredible pace, it really does raise a vital,
compelling question for all of us looking ahead, doesn't it?
Speaker 1 (37:25):
Yeah, what's the big takeaway thought?
Speaker 2 (37:26):
As robots keep evolving like this, taking on roles we
once thought were impossible, for machines, becoming more capable, more ubiquitous,
we have to seriously consider what completely new opportunities and
let's be honest, what new challenges will emerge when these
intelligent companions become truly integrated into our society.
Speaker 1 (37:46):
How does it change things fundamentally?
Speaker 2 (37:48):
How will our understanding of collaboration, of efficiency, even of creativity?
How will all that continue to be redefined maybe expanded
by this deepening partnership reforming with these incredible machines.
Speaker 1 (38:00):
It's happening now.
Speaker 2 (38:01):
It's a future that's really not science fiction anymore. It's
becoming science fact, unfolding right now, right before our eyes,
and it kind of invites all of us to watch
closely and be part of this astonishing evolution.
Speaker 1 (38:15):
Well, we hope this deep dive has given you that
powerful shortcut to being well informed, maybe spark some aha
moments for you, or even some new questions about the
incredible possibilities ahead.
Speaker 2 (38:26):
There's certainly a lot to think about.
Speaker 1 (38:27):
The world of robotics is moving incredibly fast, and we'll
definitely be here tracking every mind blowing moment, every breakthrough.
So keep exploring, keep questioning, and definitely keep an eye
on what's coming next
Speaker 2 (38:39):
Because the future is arriving faster than you think.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
New Heights with Jason & Travis Kelce
Football’s funniest family duo — Jason Kelce of the Philadelphia Eagles and Travis Kelce of the Kansas City Chiefs — team up to provide next-level access to life in the league as it unfolds. The two brothers and Super Bowl champions drop weekly insights about the weekly slate of games and share their INSIDE perspectives on trending NFL news and sports headlines. They also endlessly rag on each other as brothers do, chat the latest in pop culture and welcome some very popular and well-known friends to chat with them. Check out new episodes every Wednesday. Follow New Heights on the Wondery App, YouTube or wherever you get your podcasts. You can listen to new episodes early and ad-free, and get exclusive content on Wondery+. Join Wondery+ in the Wondery App, Apple Podcasts or Spotify. And join our new membership for a unique fan experience by going to the New Heights YouTube channel now!