Jack 🤖's banner

Jack 🤖

@JacklouisP • 12,829 subscribers

10 years of Robot-Maxxing. I invest in robotics and physical AI. Posts are for entertainment & not investment advice.

Shorts

Nature's own motor! Bacterial chemotaxis are driven by a bidirectional flagellar motor which can rotate at different rates. New research by Prash Singh highlights how the motor transmits torque in different directions Nature once again leads the way. Who will replicate?

Nature's own motor! Bacterial chemotaxis are driven by a bidirectional flagellar motor which can rotate at different rates. New research by Prash Singh highlights how the motor transmits torque in different directions Nature once again leads the way. Who will replicate?

619,068 次观看

Thread: How China bought Germany's robotics crown jewel Quick story: I recently called Kuka German in a thread and got called out - "Kuka is Chinese now." This sent me down a rabbit hole. What I found was straight out of Succession, and it reshaped how the West thinks about tech sovereignty. This is the story of how a Chinese appliance company bought Kuka & kick-started China's robot dominance. ⬇️

Thread: How China bought Germany's robotics crown jewel Quick story: I recently called Kuka German in a thread and got called out - "Kuka is Chinese now." This sent me down a rabbit hole. What I found was straight out of Succession, and it reshaped how the West thinks about tech sovereignty. This is the story of how a Chinese appliance company bought Kuka & kick-started China's robot dominance. ⬇️

1,312,115 次观看

Traditional factory automation follows three rules: 1. Minimise degrees of freedom 2. Minimise actuation 3. Minimise the requirement for feedback Let geometry and gravity do the hard work. This is a step feeder. Reciprocating plates push parts upward, one step at a time. Parts that sit stable on the ledge rise to the top. Parts that don't? They tip off and fall back into the hopper. A chain conveyor carries the parts along and a simple tool knocks off any that are misaligned. No robot. No vision. No gripper. Just continuous circulation until every part finds its perfect orientation. Bowl feeders, gravity slides, step feeders, escapements—all built on the same principle. Constrain movement. Let shape and weight do the sorting. Modern robotics throws sensors, actuators, and computing at every problem. Sometimes the smarter move is less.

Traditional factory automation follows three rules: 1. Minimise degrees of freedom 2. Minimise actuation 3. Minimise the requirement for feedback Let geometry and gravity do the hard work. This is a step feeder. Reciprocating plates push parts upward, one step at a time. Parts that sit stable on the ledge rise to the top. Parts that don't? They tip off and fall back into the hopper. A chain conveyor carries the parts along and a simple tool knocks off any that are misaligned. No robot. No vision. No gripper. Just continuous circulation until every part finds its perfect orientation. Bowl feeders, gravity slides, step feeders, escapements—all built on the same principle. Constrain movement. Let shape and weight do the sorting. Modern robotics throws sensors, actuators, and computing at every problem. Sometimes the smarter move is less.

467,244 次观看

The vibratory bowl feeder. Patented in 1950. Here is the physics behind this ubiquitous tool. It solves a universal factory problem: you have a bin of randomly oriented parts and need them single-file, perfectly aligned, feeding into the next machine. No vision. No sensors. No code. Just physics. The bowl vibrates with an asymmetric waveform - part vertical, part rotational. During the slow phase, static friction grips the part. During the fast phase, the part slips or micro-jumps. Net result: parts climb the spiral. At the top, geometry takes over. Slots, ledges, and narrowing tracks are machined for one specific part shape. Wrong orientation? Fall back in. Correct? Exit single-file. Every bowl is custom tooled. Change the part, change the bowl. Inflexible? Completely. But at high volume, nothing beats it on cost, speed, or reliability. Running 24/7 since 1950.

The vibratory bowl feeder. Patented in 1950. Here is the physics behind this ubiquitous tool. It solves a universal factory problem: you have a bin of randomly oriented parts and need them single-file, perfectly aligned, feeding into the next machine. No vision. No sensors. No code. Just physics. The bowl vibrates with an asymmetric waveform - part vertical, part rotational. During the slow phase, static friction grips the part. During the fast phase, the part slips or micro-jumps. Net result: parts climb the spiral. At the top, geometry takes over. Slots, ledges, and narrowing tracks are machined for one specific part shape. Wrong orientation? Fall back in. Correct? Exit single-file. Every bowl is custom tooled. Change the part, change the bowl. Inflexible? Completely. But at high volume, nothing beats it on cost, speed, or reliability. Running 24/7 since 1950.

379,262 次观看

Humanoids = stacked actuators. Pure and simple. If you don't understand actuators, you don't understand robots. Actuators are not solved: too heavy, too hot, too inefficient. This is a mega thread on actuator design for humanoid robots, directly from the mind of Humanoid Scott Bookmark it for later 👇

Humanoids = stacked actuators. Pure and simple. If you don't understand actuators, you don't understand robots. Actuators are not solved: too heavy, too hot, too inefficient. This is a mega thread on actuator design for humanoid robots, directly from the mind of Humanoid Scott Bookmark it for later 👇

245,119 次观看

Is this Poka Yoke or just good tooling design?

Is this Poka Yoke or just good tooling design?

231,661 次观看

First rule of automation - make your system as simple as possible but no simpler. The tending 'robot' of this wire bender is made up of one linear and one rotary actuator. Not pretty, not flashy, but if it gets the job done 99.999999% of the time with minimal maintenance required - Job is a good un.

First rule of automation - make your system as simple as possible but no simpler. The tending 'robot' of this wire bender is made up of one linear and one rotary actuator. Not pretty, not flashy, but if it gets the job done 99.999999% of the time with minimal maintenance required - Job is a good un.

227,050 次观看

What can robotics learn from watch design? - Hundreds of parts packed into a coin-sized case, - keeping time within ±20 seconds daily - enduring decades of use Mechanical design at its most optimal. Let's explore the mechanisms that make watches tick ⬇️

What can robotics learn from watch design? - Hundreds of parts packed into a coin-sized case, - keeping time within ±20 seconds daily - enduring decades of use Mechanical design at its most optimal. Let's explore the mechanisms that make watches tick ⬇️

198,117 次观看

5-axis CNC machining – but different... Most 5-axis machines use serial kinematics: stack a rotary A-axis on top of a rotary B-axis, mount that on linear X/Y/Z stages. Each axis carries the weight of everything after it. Heavy and slow. 🤖 The Sprint Z3 uses parallel kinematics: three linear drives working together to create BOTH the Z-axis motion AND the A/B tilt simultaneously. No stacked rotary joints. No heavy tilting head carrying a motor on top of another motor. The entire barrel-shaped headstock moves as one unit vertically in the column, while the three internal drives handle the spindle positioning and orientation. ⚙️ How it works: Three linear actuators are arranged in a circle, all connected to the spindle platform through pivot arms. - Extend all three equally → spindle moves straight up/down (Z-axis) - Extend them different amounts → spindle tilts (A/B axes) It's using geometry instead of stacked rotary joints to create tilt 🤔 Why this is useful: Way less moving mass than serial kinematics. Less mass = faster accelerations, higher precision, better surface finish. For high-speed machining, dynamics matter more than workspace. Question: Is this a good design or overcomplicated❓

5-axis CNC machining – but different... Most 5-axis machines use serial kinematics: stack a rotary A-axis on top of a rotary B-axis, mount that on linear X/Y/Z stages. Each axis carries the weight of everything after it. Heavy and slow. 🤖 The Sprint Z3 uses parallel kinematics: three linear drives working together to create BOTH the Z-axis motion AND the A/B tilt simultaneously. No stacked rotary joints. No heavy tilting head carrying a motor on top of another motor. The entire barrel-shaped headstock moves as one unit vertically in the column, while the three internal drives handle the spindle positioning and orientation. ⚙️ How it works: Three linear actuators are arranged in a circle, all connected to the spindle platform through pivot arms. - Extend all three equally → spindle moves straight up/down (Z-axis) - Extend them different amounts → spindle tilts (A/B axes) It's using geometry instead of stacked rotary joints to create tilt 🤔 Why this is useful: Way less moving mass than serial kinematics. Less mass = faster accelerations, higher precision, better surface finish. For high-speed machining, dynamics matter more than workspace. Question: Is this a good design or overcomplicated❓

118,947 次观看

Something very satisfying about the way this system opens bags of bulk plastic powder. Not sure if it's the slice or the wiggle

Something very satisfying about the way this system opens bags of bulk plastic powder. Not sure if it's the slice or the wiggle

102,213 次观看

Robots Making Robots in China 🤖 Kuka's Shunde factory produces one industrial robot every 30 minutes, with plans to reduce this to one minute per unit. The facility outputs 30,000 robots annually for industries from automotive to medical applications. The German company's localization strategy has paid off: sourcing 80-90% of components from southern China (versus previously from eastern China) cut production costs by one-third and delivery times from 2-3 months to 2-3 weeks. Recent growth has been driven by Chinese EV manufacturers expanding overseas and electronics companies upgrading production lines. China now represents 25% of Kuka's $4.1B annual revenue.

Robots Making Robots in China 🤖 Kuka's Shunde factory produces one industrial robot every 30 minutes, with plans to reduce this to one minute per unit. The facility outputs 30,000 robots annually for industries from automotive to medical applications. The German company's localization strategy has paid off: sourcing 80-90% of components from southern China (versus previously from eastern China) cut production costs by one-third and delivery times from 2-3 months to 2-3 weeks. Recent growth has been driven by Chinese EV manufacturers expanding overseas and electronics companies upgrading production lines. China now represents 25% of Kuka's $4.1B annual revenue.

66,920 次观看

No robots. No vision. No AI. Just a vibratory bowl feeder, a 2-axis pick-and-place, a turntable, and a gravity slide. This is the automation that actually runs the world. Bowl feeders have been orienting parts since the 1950s. Vibration and geometry—nothing else. Parts walk up a spiral track and fall into line. Pneumatic cylinders move on two axes—up/down and rotate. Hard stops set the positions. No encoders. No feedback loops. Just air pressure hitting a mechanical limit. Air jets blow parts down the slides. Gravity gets a helping hand from a tiny pulse of air to keep the flow consistent. Inflexible? Absolutely. One product, one machine. But for high-volume production, nothing beats it on cost, speed, or reliability. The unsexy backbone of manufacturing.

No robots. No vision. No AI. Just a vibratory bowl feeder, a 2-axis pick-and-place, a turntable, and a gravity slide. This is the automation that actually runs the world. Bowl feeders have been orienting parts since the 1950s. Vibration and geometry—nothing else. Parts walk up a spiral track and fall into line. Pneumatic cylinders move on two axes—up/down and rotate. Hard stops set the positions. No encoders. No feedback loops. Just air pressure hitting a mechanical limit. Air jets blow parts down the slides. Gravity gets a helping hand from a tiny pulse of air to keep the flow consistent. Inflexible? Absolutely. One product, one machine. But for high-volume production, nothing beats it on cost, speed, or reliability. The unsexy backbone of manufacturing.

38,962 次观看

This is why robots haven't replaced humans in brake line fabrication: - Every vehicle model has unique routing through cluttered underbodies; a skilled worker reads the geometry and adjusts in real time - Per-SKU volumes are low, so setup/programming costs never pay back - The manual jig is low cost; the human brings the compute, sensing, and adaptability for free When variance is high and tooling is cheap, people are still the best option. What would it take to automate this?

This is why robots haven't replaced humans in brake line fabrication: - Every vehicle model has unique routing through cluttered underbodies; a skilled worker reads the geometry and adjusts in real time - Per-SKU volumes are low, so setup/programming costs never pay back - The manual jig is low cost; the human brings the compute, sensing, and adaptability for free When variance is high and tooling is cheap, people are still the best option. What would it take to automate this?

20,990 次观看

RoboBallet: End of Manual Robot Programming Google DeepMind/Intrinsic/UCL just cracked multi-robot coordination. Published in Science Robotics. 🤖What it does: – 8 robots, 40 tasks, zero collisions - all planned in seconds – 25% better motion plans than expert-designed solutions – 60% faster task completion when scaling from 4 to 8 robots – Works on factory layouts, it's never seen 👩‍💻 The tech: – Graph Neural Networks map robots/obstacles as connected nodes – Reinforcement learning trains on millions of scenarios – No manual programming required - just give it CAD files and tasks ❓Question: The bottleneck to robot collaboration has always been manual prototyping. Now that we can automate the automation what will this unlock?

RoboBallet: End of Manual Robot Programming Google DeepMind/Intrinsic/UCL just cracked multi-robot coordination. Published in Science Robotics. 🤖What it does: – 8 robots, 40 tasks, zero collisions - all planned in seconds – 25% better motion plans than expert-designed solutions – 60% faster task completion when scaling from 4 to 8 robots – Works on factory layouts, it's never seen 👩‍💻 The tech: – Graph Neural Networks map robots/obstacles as connected nodes – Reinforcement learning trains on millions of scenarios – No manual programming required - just give it CAD files and tasks ❓Question: The bottleneck to robot collaboration has always been manual prototyping. Now that we can automate the automation what will this unlock?

41,633 次观看

Cams ARE the original "programmable" automation - these mechanical marvels orchestrate complex sequences without a single line of code. By carefully profiling cam shapes, engineers encode timing, positioning, and sequencing directly into metal. What makes cams brilliant for automation: - Deterministic timing - no software, no edge cases - Self-synchronising - no syncing clocks of different systems - Incredibly durable - cam-driven machines run for decades - Simple maintenance - mechanics can "read" the program by looking at cam profiles The cam followers, springs, and linkages in machines like this execute intricate routines - bending, cutting, feeding, and positioning with clockwork precision. Each bump and valley on those rotating cams tells a different part of the machine exactly when to move. With all the dancing robots, we sometimes forget how high-tech old-school automation is.

Cams ARE the original "programmable" automation - these mechanical marvels orchestrate complex sequences without a single line of code. By carefully profiling cam shapes, engineers encode timing, positioning, and sequencing directly into metal. What makes cams brilliant for automation: - Deterministic timing - no software, no edge cases - Self-synchronising - no syncing clocks of different systems - Incredibly durable - cam-driven machines run for decades - Simple maintenance - mechanics can "read" the program by looking at cam profiles The cam followers, springs, and linkages in machines like this execute intricate routines - bending, cutting, feeding, and positioning with clockwork precision. Each bump and valley on those rotating cams tells a different part of the machine exactly when to move. With all the dancing robots, we sometimes forget how high-tech old-school automation is.

40,142 次观看

Just discovered Dekal via Timothée Peter's video modeling a spherical five-bar linkage in code. What makes it interesting: > Code-based kinematic design with a differentiable simulation engine > Describe mechanisms in code, instantly get workspace mapping, reachability analysis, and static load testing across every joint > Idea to kinematic validation ASAP; then push to CAD Its a cool project built by a mechanical engineer frustrated with iteration speed in existing tools.

Just discovered Dekal via Timothée Peter's video modeling a spherical five-bar linkage in code. What makes it interesting: > Code-based kinematic design with a differentiable simulation engine > Describe mechanisms in code, instantly get workspace mapping, reachability analysis, and static load testing across every joint > Idea to kinematic validation ASAP; then push to CAD Its a cool project built by a mechanical engineer frustrated with iteration speed in existing tools.

14,532 次观看

The actuator dilemma that's plagued industrial robotics for 50 years: Electric motors are most efficient at high speeds (5,000-20,000 RPM). But robot joints need slow, controlled movement (~60 RPM max). This forces massive gear reductions (50:1 to 320:1) - and that's where everything breaks down.

The actuator dilemma that's plagued industrial robotics for 50 years: Electric motors are most efficient at high speeds (5,000-20,000 RPM). But robot joints need slow, controlled movement (~60 RPM max). This forces massive gear reductions (50:1 to 320:1) - and that's where everything breaks down.

24,821 次观看

Cartwheel Robotics is one of my favourite humanoids... they're building for homes, not warehouses – focusing on intangible 'companionship' rather than quantifiable 'work.' 🤖 This lowers the bar: It's not measured by productivity or ROI. No brutal cost-per-task calculations. No frustration when it fails at a job. Success is simply: do people like it? Does it make them feel something? 🎨 The design choice: Deliberately cartoon-styled to avoid the uncanny valley. Friendly and approachable, not creepy. Warmth and personality over capability. 💣 The challenge: Making something people genuinely connect with is can be harder than optimising for tasks – there's no playbook, no clear definition of done. But founder Scott LaValley (ex-Boston Dynamics, Disney – led Baby Groot, worked on Atlas) has exactly the right background. He knows how to make robots that resonate. 🎯 If they get this right, there could be a bigger moat than approaches that focus on doing 'work'... This is Disney's strategy. What do you think, is this the right humanoid strategy for our current tech?

Cartwheel Robotics is one of my favourite humanoids... they're building for homes, not warehouses – focusing on intangible 'companionship' rather than quantifiable 'work.' 🤖 This lowers the bar: It's not measured by productivity or ROI. No brutal cost-per-task calculations. No frustration when it fails at a job. Success is simply: do people like it? Does it make them feel something? 🎨 The design choice: Deliberately cartoon-styled to avoid the uncanny valley. Friendly and approachable, not creepy. Warmth and personality over capability. 💣 The challenge: Making something people genuinely connect with is can be harder than optimising for tasks – there's no playbook, no clear definition of done. But founder Scott LaValley (ex-Boston Dynamics, Disney – led Baby Groot, worked on Atlas) has exactly the right background. He knows how to make robots that resonate. 🎯 If they get this right, there could be a bigger moat than approaches that focus on doing 'work'... This is Disney's strategy. What do you think, is this the right humanoid strategy for our current tech?

11,774 次观看

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Here is a human assembling a BMW engine by hand. Hundreds of components. Micron tolerances. Dozens of hours. So what's better here—a human or a robot❓ 🤖 Robots win at: - Repeatability. Same torque, same angle, every time. - Speed on defined tasks. Honda automates piston installation at 9,000/day. - No fatigue. Shift 47 is identical to shift 1. 👨‍🏭 Humans win at: - Feeling what they can't see. Half of engine assembly is blind. - Force intuition. Knowing "seated" from "about to crack." - Compliant manipulation. That instinctive wiggle to get a part to drop in. - Handling variance. Mercedes pulled robots off the S-Class line because they couldn't adapt to option variability. - Deformables. Wiring harnesses, gaskets, hoses—nothing predictable. - Recovery. When something's wrong, humans adapt. Robots halt. ✅ Right now, the answer is both. Robots handle the repetitive subassemblies. Humans handle everything that requires judgment, feeling, and adaptation. The day a humanoid can do this end-to-end is the day manipulation is solved.

Here is a human assembling a BMW engine by hand. Hundreds of components. Micron tolerances. Dozens of hours. So what's better here—a human or a robot❓ 🤖 Robots win at: - Repeatability. Same torque, same angle, every time. - Speed on defined tasks. Honda automates piston installation at 9,000/day. - No fatigue. Shift 47 is identical to shift 1. 👨‍🏭 Humans win at: - Feeling what they can't see. Half of engine assembly is blind. - Force intuition. Knowing "seated" from "about to crack." - Compliant manipulation. That instinctive wiggle to get a part to drop in. - Handling variance. Mercedes pulled robots off the S-Class line because they couldn't adapt to option variability. - Deformables. Wiring harnesses, gaskets, hoses—nothing predictable. - Recovery. When something's wrong, humans adapt. Robots halt. ✅ Right now, the answer is both. Robots handle the repetitive subassemblies. Humans handle everything that requires judgment, feeling, and adaptation. The day a humanoid can do this end-to-end is the day manipulation is solved.

379,830 次观看 • 6 个月前

📍 Robot localisation is the deca-billion-dollar problem no one talks about. Every mobile robot needs to know its location to within 2-5 cm. Sounds simple? It's not. Current solutions force an impossible choice: cheap but inflexible, or smart but fragile. This fundamental trade-off is blocking the mass deployment of robots. Here's why - and how nature might solve it 🧵

📍 Robot localisation is the deca-billion-dollar problem no one talks about. Every mobile robot needs to know its location to within 2-5 cm. Sounds simple? It's not. Current solutions force an impossible choice: cheap but inflexible, or smart but fragile. This fundamental trade-off is blocking the mass deployment of robots. Here's why - and how nature might solve it 🧵

400,070 次观看 • 10 个月前

Should humanoids have tails? Current legged robots struggle with center of mass control and quick stabilization. Adding more DOF to legs gets complicated fast. Nature already solved this: Cheetahs, kangaroos, dinosaurs – they use tails for dynamic balance, sharp turns, and recovery from unexpected shifts. The robotics insight? Let the tail handle it. Virginia Tech built serpentine tails that compensate in real-time – so bipeds and quadrupeds can handle ledges, lateral shifts, and turns without rewriting their entire gait. Simple solution. Obvious in hindsight. 🦖🤖

Should humanoids have tails? Current legged robots struggle with center of mass control and quick stabilization. Adding more DOF to legs gets complicated fast. Nature already solved this: Cheetahs, kangaroos, dinosaurs – they use tails for dynamic balance, sharp turns, and recovery from unexpected shifts. The robotics insight? Let the tail handle it. Virginia Tech built serpentine tails that compensate in real-time – so bipeds and quadrupeds can handle ledges, lateral shifts, and turns without rewriting their entire gait. Simple solution. Obvious in hindsight. 🦖🤖

277,571 次观看 • 7 个月前

Another trade show demo - this time showing off how perfectly in sync these motors are. Thoughts on why you would need this level of control?

Another trade show demo - this time showing off how perfectly in sync these motors are. Thoughts on why you would need this level of control?

314,167 次观看 • 9 个月前

Actuators eat up 30-50% of a humanoid robot's BOM cost. Mass adoption? Not at these prices. This is one of the most persistent challenges in robotics, but IMSystems may have cracked the code with a novel drive. To understand it, we need to dive into the world of drives & transmission, with all their demands and trade-offs. Shoutout to Marwa ElDiwiny and Humanoid Scott for the original video.

Actuators eat up 30-50% of a humanoid robot's BOM cost. Mass adoption? Not at these prices. This is one of the most persistent challenges in robotics, but IMSystems may have cracked the code with a novel drive. To understand it, we need to dive into the world of drives & transmission, with all their demands and trade-offs. Shoutout to Marwa ElDiwiny and Humanoid Scott for the original video.

358,349 次观看 • 11 个月前

Check out Robuild's autonomous construction robots They recently completed a 500 ft² LVP flooring installation, with 600 install planned by 2026. A few tech details: • Labour shortages & cost savings are driving developers to look at robotics • The system works in busy, varied houses and uses onboard vision to align and place planks without pre-mapping. • It's built on ROS, using NVIDIA Omniverse, Isaac Gym, and custom PID for precision.

Check out Robuild's autonomous construction robots They recently completed a 500 ft² LVP flooring installation, with 600 install planned by 2026. A few tech details: • Labour shortages & cost savings are driving developers to look at robotics • The system works in busy, varied houses and uses onboard vision to align and place planks without pre-mapping. • It's built on ROS, using NVIDIA Omniverse, Isaac Gym, and custom PID for precision.

223,954 次观看 • 11 个月前

This machinetakes flat aluminum sheet and produces complex 3D products - up to 110 parts per hour. ❓The question everyone asks: how does it rotate the sheet? Look at the center of the bed. There's a channel running through it where a vacuum suction manipulator grips the sheet from underneath. A servo-driven rotator spins it—90°, 180°, full 360°—to present each edge to the bending blades in sequence. The sheet never leaves the horizontal plane. No external robot arm. No regripping. Just one continuous cycle from flat blank to finished part. This is what separates panel bending centers from traditional press brakes with bolt-on automation.

This machinetakes flat aluminum sheet and produces complex 3D products - up to 110 parts per hour. ❓The question everyone asks: how does it rotate the sheet? Look at the center of the bed. There's a channel running through it where a vacuum suction manipulator grips the sheet from underneath. A servo-driven rotator spins it—90°, 180°, full 360°—to present each edge to the bending blades in sequence. The sheet never leaves the horizontal plane. No external robot arm. No regripping. Just one continuous cycle from flat blank to finished part. This is what separates panel bending centers from traditional press brakes with bolt-on automation.

97,903 次观看 • 6 个月前

AI vision models perform well on standard benchmarks like ImageNet, often matching or exceeding human accuracy. But here's what they don't tell you: Move an image by ONE PIXEL and these same models have a meltdown 40% of the time. This brittleness is breaking real systems.

AI vision models perform well on standard benchmarks like ImageNet, often matching or exceeding human accuracy. But here's what they don't tell you: Move an image by ONE PIXEL and these same models have a meltdown 40% of the time. This brittleness is breaking real systems.

122,489 次观看 • 9 个月前

In 1980, Marc Raibert started the Leg Lab at Carnegie Mellon with a fundamental question: ⚖️ What if balance comes from movement, not stillness? At the time, robots were bolted down. Motion meant precision and repeatability. The idea was to build deterministic robots that found stability in their predictability. Their key insight - nature doesn't work that way, why should robots? ⚙️ What they built: Robots that bounced and stumbled but stayed upright through dynamic motion, not static stability. It looked awkward. It wasn't immediately useful. But it worked. Every hop, every stumble, every recovery taught them how dynamic balance actually works. 🎯 The result: Raibert spun out Boston Dynamics (1992), that DNA came with him. BigDog, Atlas, Spot – all descendants of those early hopping machines. Every humanoid today: Figure, Tesla, Agility - uses principles discovered in the Leg Lab. Dynamic stability. Active balance. Movement as intelligence.

In 1980, Marc Raibert started the Leg Lab at Carnegie Mellon with a fundamental question: ⚖️ What if balance comes from movement, not stillness? At the time, robots were bolted down. Motion meant precision and repeatability. The idea was to build deterministic robots that found stability in their predictability. Their key insight - nature doesn't work that way, why should robots? ⚙️ What they built: Robots that bounced and stumbled but stayed upright through dynamic motion, not static stability. It looked awkward. It wasn't immediately useful. But it worked. Every hop, every stumble, every recovery taught them how dynamic balance actually works. 🎯 The result: Raibert spun out Boston Dynamics (1992), that DNA came with him. BigDog, Atlas, Spot – all descendants of those early hopping machines. Every humanoid today: Figure, Tesla, Agility - uses principles discovered in the Leg Lab. Dynamic stability. Active balance. Movement as intelligence.

87,993 次观看 • 6 个月前

Welding robots. The original use case for industrial automation—first deployed in the 1960s at GM. 60 years later, they dominate automotive body shops. Hundreds of robots in perfect sync, laying thousands of spot welds per vehicle. But step outside high-volume production and they're rare. Most fab shops still weld by hand. Why? Every new part needs skilled programming. Path planning. Torch angle. Travel speed. Wire feed. Gas flow. All tuned for each weld on each component. The economics only work if you're making thousands of the same thing. High-volume? Solved. High-mix, low-volume? Traditional welding robots fail here... but watch this space a few start ups are starting to fill the gap.

Welding robots. The original use case for industrial automation—first deployed in the 1960s at GM. 60 years later, they dominate automotive body shops. Hundreds of robots in perfect sync, laying thousands of spot welds per vehicle. But step outside high-volume production and they're rare. Most fab shops still weld by hand. Why? Every new part needs skilled programming. Path planning. Torch angle. Travel speed. Wire feed. Gas flow. All tuned for each weld on each component. The economics only work if you're making thousands of the same thing. High-volume? Solved. High-mix, low-volume? Traditional welding robots fail here... but watch this space a few start ups are starting to fill the gap.

78,204 次观看 • 6 个月前

Fancy a self-driving bike? The design for this one is open sourced & available on Github ⏬ ⏬ ⏬

Fancy a self-driving bike? The design for this one is open sourced & available on Github ⏬ ⏬ ⏬

26,741 次观看 • 3 个月前

OpenArm has been released - A fully open-source bimanual robot arm built for physical AI. Some stats: – 7 DoF – 633mm reach – 6kg payload/arm All in BOM cost of $6.5k. Thanks for sharing - Enactic Hiroto Yamamoto | 山本泰豊

OpenArm has been released - A fully open-source bimanual robot arm built for physical AI. Some stats: – 7 DoF – 633mm reach – 6kg payload/arm All in BOM cost of $6.5k. Thanks for sharing - Enactic Hiroto Yamamoto | 山本泰豊

63,595 次观看 • 10 个月前

☕ Coffee extraction: a $100B optimisation problem disguised as a 30-second routine. Every shot balances 6 variables and 1000+ molecules. One parameter off? Expensive disappointment. You'd think this has nothing to do with robotics but you would be wrong, lets dig in ⬇️

☕ Coffee extraction: a $100B optimisation problem disguised as a 30-second routine. Every shot balances 6 variables and 1000+ molecules. One parameter off? Expensive disappointment. You'd think this has nothing to do with robotics but you would be wrong, lets dig in ⬇️

55,239 次观看 • 9 个月前

🤖 Ultrafast Insect-Scale Robotics Check out this new robot design published in Nature: 🐞 Insect Inspiration: • Mimics high-speed locomotion of insects like tiger beetles • Adopts running gait similar to cockroaches • Achieves agile turning comparable to honey bees Key Features: • Size: 2cm long, 1760mg weight • Speed: 17.5 body lengths (BL) /second • Turning acceleration: 65.4 BL/s² • Wireless controlled 🛠️ How It Works: 1. Electromagnetic Actuation: • Two independent actuators control front legs • Generates high-frequency vibrations (up to 220 Hz) 2. Bouncing Mechanism: • Front legs longer than rear, creating body tilt • Periodic impacts between legs and ground produce bouncing 3. Gait Cycle: • Includes squatting, bouncing, aerial, and landing phases • Complementary combination of bounce length and frequency 4. Wireless Control: • Onboard circuit with Bluetooth for remote operation • Two independent channels for precise movement control Applications: 🔍 Search and rescue in confined spaces 🛩️ Internal inspections of turbofan engines 🚁 Drone-assisted deployment for extended reach

🤖 Ultrafast Insect-Scale Robotics Check out this new robot design published in Nature: 🐞 Insect Inspiration: • Mimics high-speed locomotion of insects like tiger beetles • Adopts running gait similar to cockroaches • Achieves agile turning comparable to honey bees Key Features: • Size: 2cm long, 1760mg weight • Speed: 17.5 body lengths (BL) /second • Turning acceleration: 65.4 BL/s² • Wireless controlled 🛠️ How It Works: 1. Electromagnetic Actuation: • Two independent actuators control front legs • Generates high-frequency vibrations (up to 220 Hz) 2. Bouncing Mechanism: • Front legs longer than rear, creating body tilt • Periodic impacts between legs and ground produce bouncing 3. Gait Cycle: • Includes squatting, bouncing, aerial, and landing phases • Complementary combination of bounce length and frequency 4. Wireless Control: • Onboard circuit with Bluetooth for remote operation • Two independent channels for precise movement control Applications: 🔍 Search and rescue in confined spaces 🛩️ Internal inspections of turbofan engines 🚁 Drone-assisted deployment for extended reach

89,282 次观看 • 1 年前

What if we're approaching robotic manipulation all wrong? Ilya Sutskever makes a fascinating point here: Current approaches throw massive amounts of data at the problem — millions of simulation steps, huge compute. And still, robots can't match human dexterity. Real-world learning of new skills? "Very out of reach." Meanwhile, humans pick up manipulation tasks almost instantly. Our dexterity is unmatched. And we do it with remarkable robustness — no reward shaping, no curriculum design, no training instability. Why the gap? Evolution. 500M years of optimisation for locomotion, vision, and manipulation. For these ancient sensorimotor tasks, maybe brute-force data isn't the answer. Maybe we should be copying the biological algorithms that already solved it. CC: Dwarkesh Patel

What if we're approaching robotic manipulation all wrong? Ilya Sutskever makes a fascinating point here: Current approaches throw massive amounts of data at the problem — millions of simulation steps, huge compute. And still, robots can't match human dexterity. Real-world learning of new skills? "Very out of reach." Meanwhile, humans pick up manipulation tasks almost instantly. Our dexterity is unmatched. And we do it with remarkable robustness — no reward shaping, no curriculum design, no training instability. Why the gap? Evolution. 500M years of optimisation for locomotion, vision, and manipulation. For these ancient sensorimotor tasks, maybe brute-force data isn't the answer. Maybe we should be copying the biological algorithms that already solved it. CC: Dwarkesh Patel

31,233 次观看 • 6 个月前

I feel like we’re missing the fact that Disney is building us actual droids

I feel like we’re missing the fact that Disney is building us actual droids

70,377 次观看 • 1 年前

I missed this when it was released a few weeks ago... probs the most unhinged crab robot out there. Love to see it. What it is: A heavy-duty construction robot that can swap tools and arms automatically during operation. Think modular platform rather than a single-purpose machine. Key specs: - 100kg payload, 4m reach, Fits in a Euro-pallet for easy transport - Autonomous tool switching via standardised interfaces - Handles placement, fastening, finishing, and inspection What makes it different: Designed from first principles for construction work, not human mimicry. The modular architecture means one robot tackles multiple tasks across different build phases without redesign.

I missed this when it was released a few weeks ago... probs the most unhinged crab robot out there. Love to see it. What it is: A heavy-duty construction robot that can swap tools and arms automatically during operation. Think modular platform rather than a single-purpose machine. Key specs: - 100kg payload, 4m reach, Fits in a Euro-pallet for easy transport - Autonomous tool switching via standardised interfaces - Handles placement, fastening, finishing, and inspection What makes it different: Designed from first principles for construction work, not human mimicry. The modular architecture means one robot tackles multiple tasks across different build phases without redesign.

29,454 次观看 • 9 个月前

WPP trialled Boston Dynamics' Atlas as a camera operator in LED volume production 3 months ago. Key capabilities: - Precise repeatability of complex camera moves - Sideways walking motions are impossible for humans - Eliminates setup time vs traditional robotic rigs - Operates safely in hazardous filming conditions Testing included automotive and F&B commercials. Directors noted similarities to working with human operators but with enhanced precision.

WPP trialled Boston Dynamics' Atlas as a camera operator in LED volume production 3 months ago. Key capabilities: - Precise repeatability of complex camera moves - Sideways walking motions are impossible for humans - Eliminates setup time vs traditional robotic rigs - Operates safely in hazardous filming conditions Testing included automotive and F&B commercials. Directors noted similarities to working with human operators but with enhanced precision.

18,228 次观看 • 10 个月前

Can a swarm of robots build a car from scratch? AMBOTS has developed a proof of concept demonstrating - Mobile 3D printers & arms collaborating to produce a car without a human in sight. Will factories of the future be built along these lines?

Can a swarm of robots build a car from scratch? AMBOTS has developed a proof of concept demonstrating - Mobile 3D printers & arms collaborating to produce a car without a human in sight. Will factories of the future be built along these lines?

23,001 次观看 • 1 年前

We've all heard of Meta's $100M job offers for AI researchers, but these are 1 in a billion. I analysed 1,359 robotics job openings to find better odds. The logic that everyone ignores: Deployment + Manufacturing + Field Operations: 20% of jobs AI/ML Development: 15.3% of jobs While everyone fights to build AI, companies desperately need people to deploy and scale them. What they're hiring for: → Robot maintenance in warehouses ($120K+) → Factory production scaling → Field troubleshooting and integration → System reliability engineering The best robotics careers might not look like "robotics careers." The future isn't in the lab—it's in making robots work at scale

We've all heard of Meta's $100M job offers for AI researchers, but these are 1 in a billion. I analysed 1,359 robotics job openings to find better odds. The logic that everyone ignores: Deployment + Manufacturing + Field Operations: 20% of jobs AI/ML Development: 15.3% of jobs While everyone fights to build AI, companies desperately need people to deploy and scale them. What they're hiring for: → Robot maintenance in warehouses ($120K+) → Factory production scaling → Field troubleshooting and integration → System reliability engineering The best robotics careers might not look like "robotics careers." The future isn't in the lab—it's in making robots work at scale

12,029 次观看 • 11 个月前

没有更多内容可加载