kore

AI believer. Exploring the digital frontier

[kore @koreshai]

A wearable robot should carry its own weight. Literally.

This looks like an exoskeleton concept, complete with a crowd filming the performance. The frame makes a convincing visual. Its contribution to the movement is harder to judge.

For a real product, a useful demonstration would compare the same everyday task with assistance on and off. Show how much effort it saves after accounting for the equipment the wearer has to carry.

[NeuralForge AI Robots @NeuralForgef]

https://t.co/RsR0u10Qhb

[kore @koreshai]

@LeoKharon the most useless robot I’ve ever seen😁

[FluxAson @FluxAson]

The video shows a group of humanoids learning to move and operate together. South Korea's military is actively testing the concept of AI + robotics as a way to partially offset the country's shrinking number of soldiers.

And here's the most interesting part: robots can potentially learn much faster than humans. Once a task is mastered in a digital environment, the trained model transfers those skills to a physical robot.

This is already moving beyond simple experiments. In 2026, South Korea's military established a specialized tactical unit to test drones and robots. And the Navy conducted its first combat experiment in which the humanoid robot PIBOT performed ship-steering functions.

South Korea is also planning to invest 2.3 trillion won (about $1.66 billion) by 2030 into the development of humanoid robots.

But the bigger picture is even more interesting. Instead of simply replacing soldiers with machines, South Korea is developing the MUM-T (Manned-Unmanned Teaming) concept, where humans and machines operate as a single combat system.

Drones scout ahead, ground robots inspect dangerous areas, and AI helps coordinate the entire system in real time.

The only question is: how long until robots like these become a normal part of the military?

[kore @koreshai]

@FluxAson if one robot figures something out and that skill can be pushed to the rest, military adoption could move a lot faster than people expect

[kore @koreshai]

The outfit says 2050. The landing says beta.

A humanoid in silver goes through a dance routine, then ends up flat on the stage. Now comes the move worth seeing: whether it can get back up and carry on. The clip ends before we get that answer.

For anyone paying to put a robot in front of an audience, recovery belongs in the audition too.

[VynqorxeAI @Vynqorxe]

https://t.co/KdL9Vejt7i

[kore @koreshai]

Imagine paying $5,000 for a robot appearance and having to rehearse the guests.

This clip looks like a concept for a future fashion expo: a silver humanoid walks the runway, waves, then interacts with 2 visitors. That last scene is where the booking gets interesting.

A runway routine has a route and a schedule. Visitors bring interruptions, unclear questions and jokes nobody rehearsed. At that hypothetical $5,000 fee, an organiser would want to know how much conversation the machine could handle before a staff member needed to step in.

The useful version would greet guests, explain the collection and hand an unfamiliar question to the right person. Give it those abilities, and the organiser has a reason to keep it working after the cameras finish filming the entrance.

The walk earns applause. The conversation earns the next booking.

[Bloott @Bloottlp1]

https://t.co/j3zgRWgyi0

[kore @koreshai]

Imagine paying $50,000 for a robot that still needs someone to pick it up.

This compilation moves from football pitches to demo rooms and a running track. Several promising starts end on the floor. The people waiting nearby suddenly become part of the operating budget.

Take a hypothetical example: a $50,000 machine needs 1 hour of human assistance per working day. Assume that time costs $30 an hour across 250 days. That adds $7,500 a year, or $37,500 over 5 years.

The purchase plus that assistance now totals $87,500, before electricity and repairs. Small interruptions can add up to a substantial staffing bill.

Those figures are illustrative; this compilation doesn’t establish any robot’s actual failure rate. It does show why the time spent resetting a machine belongs in the calculation of how much work it saves.

Somewhere in the automation budget, there’s still a timesheet.

[HodlReaper @HodlReaper]

https://t.co/mWyHk0K6Oj

[kore @koreshai]

@0xQqwzex the best part is how fast everyone forgets it’s a robot and just starts having fun with it)

[kore @koreshai]

@0xGenpix In any case, this will give the company a significant advantage over other companies

[kore @koreshai]

A HUMANOID COULD BECOME THE MOST PHOTOGRAPHED THING A BRAND OWNS

In this imagined reveal, the phones are raised before the robot starts dancing. The exposed mechanics make the scene easy to understand, and a few movements give the crowd something to keep filming.

Picture an events company renting that character for a store opening on Friday and a product launch on Saturday. Each client gets an appearance designed to draw people over. The operator handles transport, setup and keeping the machine running. Repeat bookings would depend on whether the appearance was worth the trouble.

That puts an interesting limit on the spectacle. A robot could attract a crowd so successfully that everyone leaves with a video of it and barely remembers the product beside it. The character would need to belong in the campaign, with something to say or demonstrate that connects the attention to the business.

The client is paying to be remembered.

[Shredder @Shred_0x]

https://t.co/DYkh71pKPf

[kore @koreshai]

@LeoKharon humo family😁

[kore @koreshai]

@0xGenpix companies are starting to sell “zero-friction companionship” as a product and a lot of people may decide that convenience beats complexity.

[Xiaopeng He @xiaopenghexpeng]

Today, we launched the world’s first automated production line for advanced general‑purpose humanoid robots, using robots to mass‑produce robots, autonomously.

This was uncharted territory, and we solved it from scratch. The commissioning of our production line means humanoid robots are now ready to scale up and step into the real world.

What touched my heartstrings most was watching our very first advanced general‑purpose humanoid robot complete assembly and walk off the line on its own. To me, this is not just about a smarter tool or a commodity. I hope it gains true generalization ability to take on dangerous, repetitive, or undesirable tasks, and ultimately make life better. Perhaps one day, robots may become our companions, friends, or part of family.

Grateful to every colleague on this momumental mission. And a warm welcome to IRON!

[kore @koreshai]

1 HUMANOID. 6 PLATES. 0 FAILS.

An 11-second clip imagines a robot waiter moving through a busy restaurant, carrying a full tray, stopping at the table and placing the food directly in front of a customer.

The interesting part isn’t the tray.

It’s everything happening around it.

People keep eating. Tables are crowded. Chairs sit in the robot’s path. Customers move without warning. Nobody is standing behind a safety line waiting for the demo to finish.

That is a completely different problem from making a humanoid carry an object across a lab.

A restaurant turns simple motion into a chain of small decisions:

don’t spill → find the table → avoid people → identify the order → place it correctly → move on

And hospitality adds another layer. A useful waiter doesn’t just transport food. It has to understand when to approach, where to stand, who ordered what, and when the best action is simply to stay out of the way.

For years, humanoid robots have been sold with spectacular abilities.

Run. Jump. Dance. Lift.

But the jobs that may actually bring them into everyday life could look much less dramatic.

Carry 6 plates.

Find table 12.

Don’t spill the fries.

Then do it again 200 times.

[Xiaopeng He @xiaopenghexpeng]

Today, we launched the world’s first automated production line for advanced general‑purpose humanoid robots, using robots to mass‑produce robots, autonomously.

This was uncharted territory, and we solved it from scratch. The commissioning of our production line means humanoid robots are now ready to scale up and step into the real world.

What touched my heartstrings most was watching our very first advanced general‑purpose humanoid robot complete assembly and walk off the line on its own. To me, this is not just about a smarter tool or a commodity. I hope it gains true generalization ability to take on dangerous, repetitive, or undesirable tasks, and ultimately make life better. Perhaps one day, robots may become our companions, friends, or part of family.

Grateful to every colleague on this momumental mission. And a warm welcome to IRON!

[kore @koreshai]

@xiaopenghexpeng that is how, in the near future, robots will replace human workers in these sectors, and the unemployment rate will rise significantly

[kore @koreshai]

A HUMANOID ROBOT IN A RUSSIAN TECH STORE APPEARS TO FIGHT BACK AFTER A CUSTOMER PUSHES IT.

The clip looks almost comedic at first.

A man approaches the robot, pushes it away, and seconds later the machine moves back toward him while staff rush in to stop it.

But the interesting part isn’t whether the robot was “angry.”

We don’t even know if the response was autonomous, pre-programmed, remotely controlled, or staged.

The interesting part is how quickly the question has changed.

A few years ago, most AI mistakes happened on a screen.

Wrong answer.

Bad image.

Broken recommendation.

Humanoid robots move the same software problem into the physical world.

Now a bad response can move arms, legs, tools, and eventually much heavier machines.

That makes reliability feel very different.

The next phase of AI safety won’t only be about what machines say.

It’ll be about what they can physically do.

[kore @koreshai]

@Abobsterina the real question isn’t whether people will bond with humanoids - they probably will. it’s whether those bonds make human relationships easier to maintain, or easier to avoid)

[kore @koreshai]

@zhngjn357700 @robixone Who knows, who knows😁

[kore @koreshai]

A HUMANOID ROBOT WAS DROPPED FROM A ROOF TO SEE WHAT HAPPENS AFTER THE IMPACT.

The footage is labeled "DROP & ROLL EVALUATION." The machine falls several meters, hits the ground, folds into the impact, rolls, then works its way back toward a stable position.

It looks brutal. That’s exactly why the test is useful.

Humanoid robotics is usually shown at its cleanest: walking across flat floors, carrying boxes, dancing on stage. But useful machines won’t spend their lives inside demos. They’ll miss steps, lose balance, get knocked sideways and end up in positions nobody would deliberately choose.

A fall compresses a huge amount of engineering into a few seconds.

The body has to survive the impact.
The joints can’t become the weak point.
Sensors have to understand where the machine ended up.
Software has to decide how to recover from a posture that may never have appeared in the original plan.

Walking is only half the problem.

FALL → ABSORB → REORIENT → RECOVER → CONTINUE

Humans do this almost instinctively. For a humanoid, every arrow represents another engineering problem that has to work outside the lab.

The impressive part isn’t that the robot can fall.

It’s that engineers are starting to treat failure as part of normal operation.

[HodlReaper @HodlReaper]

https://t.co/mWyHk0K6Oj

[kore @koreshai]

@robixone one successful recovery proves the concept)