Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

🚨 #UIST2024 Paper: TorqueCapsules 💊 We introduce a novel prototyping toolkit that can exert force, jump, or flip with encapsulated flywheel actuators. The device makes flywheel actuators easy/accessible to control and customize. The University of Chicago AxLab - Actuated Experience Lab Willa Yunqi Yang 🧵→

Ken Nakagaki

7,350 subscribers

13,947 views • 1 year ago •via X (Twitter)

Science & Technology Health & Wellness Education #UIST2024

Anya Rossi• Live Now

Private livecam show

7 Comments

Ken Nakagaki1 year ago

The device is highly customizable and reconfigurable with swappable lids/flywheels. Its self-contained design makes it easy to attach or embed into existing objects. Our accompanying control GUI makes it easy to control and design the actuation of the device.

Ken Nakagaki1 year ago

While flywheel actuators have diverse and unique capabilities being used across haptics/robotics, it has been difficult to prototype and employ them. TorqueCapsules is designed to tackle this challenge by making people easily use flywheels' unique and diverse capabilities.

Ken Nakagaki1 year ago

As such, TorqueCapsules can be employed for diverse applications: - actuating everyday objects 📱 - making your plush toys jump and flip 🧸 - or building custom haptic controllers 🎮

Ken Nakagaki1 year ago

Title: TorqueCapsules: Fully-Encapsulated Flywheel Actuation Modules for Designing and Prototyping Movement-Based and Kinesthetic Interaction Authors: Willa Yunqi Yang @willayangy, Yifan Zou @yifan_zou_bot, Jingle Hung, Raouf Abujaber, Ken Nakagaki Paper:

Ken Nakagaki1 year ago

Project Web: Full Video:

Ken Nakagaki1 year ago

This was led by Willa @willayangy, one of AxLab's first PhDs. She aspires to enrich everyday environments with actuation - many more works to come ;) A big shout-out to her co-authors, especially Yifan @yifan_zou_bot, who is applying for grad school/PhD! Hire her for PhD - she is a fantastic researcher/artist/engineer!

tomoznow1 year ago

@UChicago @AxLab_UChicago @willayangy Brilliant. Hope it comes to market. I'd like an easy to config actuator to use with webhooks. So, "press button in my web app" > "moves/rotates physical thing". And integration with home automation. PIR sensor > rules engine > actuator (ring physical bell/other cool stuff.)

Related Videos

🚨 #UIST2024 Paper: CARDinality🃏 'What if ROBOTs🤖 could take the form of CARDs🃏? ' We built thin, interactive card-shaped robots that can freely locomote on tables & provide haptics in hands, both using the same actuation source, vibration. AxLab - Actuated Experience Lab The University of Chicago 🧵→

🚨 #UIST2024 Paper: CARDinality🃏 'What if ROBOTs🤖 could take the form of CARDs🃏? ' We built thin, interactive card-shaped robots that can freely locomote on tables & provide haptics in hands, both using the same actuation source, vibration. AxLab - Actuated Experience Lab The University of Chicago 🧵→

Ken Nakagaki

28,802 views • 1 year ago

🚨 #UIST2024 Paper: SHAPE-IT🪄 SHAPE-IT enables users to freely 'tell' the shape display to transform by only giving text instructions using LLMs. It offers an experience like that Sci-Fi Clay that can shape-shift whatever you say it to be✨🪄 AxLab - Actuated Experience Lab The University of Chicago →🧵

🚨 #UIST2024 Paper: SHAPE-IT🪄 SHAPE-IT enables users to freely 'tell' the shape display to transform by only giving text instructions using LLMs. It offers an experience like that Sci-Fi Clay that can shape-shift whatever you say it to be✨🪄 AxLab - Actuated Experience Lab The University of Chicago →🧵

Ken Nakagaki

15,561 views • 1 year ago

At FNT, our main focus is making crypto accessible, secure, and easy to manage, whether you're a beginner or an expert. So we designed our ecosystem in a way to give you that seamless experience and help you navigate the crypto world with ease! Read more below: 🧵

At FNT, our main focus is making crypto accessible, secure, and easy to manage, whether you're a beginner or an expert. So we designed our ecosystem in a way to give you that seamless experience and help you navigate the crypto world with ease! Read more below: 🧵

FNT Crypto

47,349 views • 1 year ago

Hands full of dexterity! 🪬 It is well known that agile, versatile hands for robots are one of the Holy Grails in robotics today. Grippers and robot hands are often underestimated, and they are the actuators behind correct and accurate picking, or placing. This robotic hand was created at the Korea University of Technology and Education. 🇰🇷 The 5-finger, 12-DoF (degrees of freedom) robotic hand features 3-DoF each for the thumb, index, and middle fingers, while the ring and little fingers share a combined 3 DoF. It uses BLDC motors with ball-screw actuators and includes a clutch mechanism to absorb impacts and prevent overload. Now, a challenge for you - guess the year! Kudos to all gripper/hand-developing companies! WE CAN'T PICK WITHOUT YOU! 🏋🏼 ~~~ ♻️ RT to help 1 robot find a new workplace.

Hands full of dexterity! 🪬 It is well known that agile, versatile hands for robots are one of the Holy Grails in robotics today. Grippers and robot hands are often underestimated, and they are the actuators behind correct and accurate picking, or placing. This robotic hand was created at the Korea University of Technology and Education. 🇰🇷 The 5-finger, 12-DoF (degrees of freedom) robotic hand features 3-DoF each for the thumb, index, and middle fingers, while the ring and little fingers share a combined 3 DoF. It uses BLDC motors with ball-screw actuators and includes a clutch mechanism to absorb impacts and prevent overload. Now, a challenge for you - guess the year! Kudos to all gripper/hand-developing companies! WE CAN'T PICK WITHOUT YOU! 🏋🏼 ~~~ ♻️ RT to help 1 robot find a new workplace.

Lukas Ziegler

33,535 views • 1 year ago

While we study the physics of "spin" and "counter-spin" as used for UFO levitation, we may wish to review current knowledge of "spin" and "counter-spin" in more well-known systems, say those used in hoverboards! In this (old) video, you can see how the fast spin of a lower flywheel with 6 N-S-N-S-N-S magnets (similar to what we have found for the lower flywheel of our UFO model) produces Lenz-inductive spin and/or rise of an aluminum disc placed above. If we use a drive-motor to "spin" that aluminum disc in one direction, while we "counter-spin" those N-S-N-S-N-S magnets in the other direction, then we get greatly enhanced repulsions due to fast RELATIVE motions! Such physics may be relevant to UFO flight, so we can study it and contemplate possibilities.🛸

While we study the physics of "spin" and "counter-spin" as used for UFO levitation, we may wish to review current knowledge of "spin" and "counter-spin" in more well-known systems, say those used in hoverboards! In this (old) video, you can see how the fast spin of a lower flywheel with 6 N-S-N-S-N-S magnets (similar to what we have found for the lower flywheel of our UFO model) produces Lenz-inductive spin and/or rise of an aluminum disc placed above. If we use a drive-motor to "spin" that aluminum disc in one direction, while we "counter-spin" those N-S-N-S-N-S magnets in the other direction, then we get greatly enhanced repulsions due to fast RELATIVE motions! Such physics may be relevant to UFO flight, so we can study it and contemplate possibilities.🛸

Red Collie (Dr. Horace Drew) scientist/inventor

52,320 views • 4 months ago

Today I finally get to play in the snow in a yard. MY YARD! AND I LOVE IT! I live in a house that love bought. We always lived in apartments because there's no wheelchair accessible houses to rent. With a LOT of help and a human friend to split it with, we finally got a house with ramps and an accessible bathroom. And today I finally got frolic and jump and run and roll in the snow! ❤️🐾❤️ Also, the leaf blower makes snoot storms and that's so fun! #Snow #DogLovers Video description for inclusivity: clips of Chesnyy the Servicerottie and her visiting lab puppy, Moon, playing in snow in her back yard

Today I finally get to play in the snow in a yard. MY YARD! AND I LOVE IT! I live in a house that love bought. We always lived in apartments because there's no wheelchair accessible houses to rent. With a LOT of help and a human friend to split it with, we finally got a house with ramps and an accessible bathroom. And today I finally got frolic and jump and run and roll in the snow! ❤️🐾❤️ Also, the leaf blower makes snoot storms and that's so fun! #Snow #DogLovers Video description for inclusivity: clips of Chesnyy the Servicerottie and her visiting lab puppy, Moon, playing in snow in her back yard

Team Servicerottie🇨🇦🐕‍🦺🦽

29,503 views • 8 months ago

The basic idea behind "UFO physics" is very simple. We can get a strong "upward" force from a fast-spinning magnet (or magnetic flywheel), because it "pushes against space" like a spinning gyroscope. Very easy, if we apply a rotating magnetic field (3 or 5-phase) to that magnet! The problem is that such an "upward" force is accompanied by an equal and opposite "downward" force by Newton's 3rd Law. So we see an increase of gravitational potential, but no loss of total weight ("up" plus "down"), unless that spinning magnet goes into "free fall". How to solve this great problem? The E.T. crop artists have suggested (many times) that we should try to change that "downward" force into another kind of force, which does NOT follow Newton's 3rd Law: namely the Lorentz force of electromagnetism. As shown in a video below, added DC current makes a disc magnet spin rapidly in one direction (by the Lorentz force), while other thin wires which supply that DC current spin in an opposite direction. We can see right away, that the mass of that disc magnet is perhaps 100 times more than the mass of many thin wires which surround it below. Yet they spin in opposite directions at approximately the same speeds! This shows that spin angular momentum (m x v x r) is NOT conserved, due to a peculiar sideways action of the Lorentz force. Does everyone understand so far? In summary, if we can change the "downward" force, for an upward spinning-magnet, into a Lorentz force, then we should have NO real "downward" force, but only a force that goes "up"! So that will be my next experiment. I have had to order some specialized parts to make the trial apparatus correctly, but will try to do it in October, after all needed parts arrive. "The simple is great" (Lao Tzu) 🛸

The basic idea behind "UFO physics" is very simple. We can get a strong "upward" force from a fast-spinning magnet (or magnetic flywheel), because it "pushes against space" like a spinning gyroscope. Very easy, if we apply a rotating magnetic field (3 or 5-phase) to that magnet! The problem is that such an "upward" force is accompanied by an equal and opposite "downward" force by Newton's 3rd Law. So we see an increase of gravitational potential, but no loss of total weight ("up" plus "down"), unless that spinning magnet goes into "free fall". How to solve this great problem? The E.T. crop artists have suggested (many times) that we should try to change that "downward" force into another kind of force, which does NOT follow Newton's 3rd Law: namely the Lorentz force of electromagnetism. As shown in a video below, added DC current makes a disc magnet spin rapidly in one direction (by the Lorentz force), while other thin wires which supply that DC current spin in an opposite direction. We can see right away, that the mass of that disc magnet is perhaps 100 times more than the mass of many thin wires which surround it below. Yet they spin in opposite directions at approximately the same speeds! This shows that spin angular momentum (m x v x r) is NOT conserved, due to a peculiar sideways action of the Lorentz force. Does everyone understand so far? In summary, if we can change the "downward" force, for an upward spinning-magnet, into a Lorentz force, then we should have NO real "downward" force, but only a force that goes "up"! So that will be my next experiment. I have had to order some specialized parts to make the trial apparatus correctly, but will try to do it in October, after all needed parts arrive. "The simple is great" (Lao Tzu) 🛸

Red Collie (Dr. Horace Drew) scientist/inventor

29,769 views • 9 months ago

A planetary roller screw looks simple, but it’s one of the key enablers for humanoid robots. It converts rotational motion into precise linear motion with high efficiency, high load capacity, and extremely smooth control. In humanoid robots, this is exactly what you need for: joint actuation linear actuators in arms/legs high-force, high-precision movement under continuous cycles Compared to ball screws, it offers better durability and load distribution—critical for robots that must operate 24/7 in the real world. Small component, huge impact. This is the kind of “hidden hardware” that makes humanoid robotics actually work.

A planetary roller screw looks simple, but it’s one of the key enablers for humanoid robots. It converts rotational motion into precise linear motion with high efficiency, high load capacity, and extremely smooth control. In humanoid robots, this is exactly what you need for: joint actuation linear actuators in arms/legs high-force, high-precision movement under continuous cycles Compared to ball screws, it offers better durability and load distribution—critical for robots that must operate 24/7 in the real world. Small component, huge impact. This is the kind of “hidden hardware” that makes humanoid robotics actually work.

Steven Cheng

69,777 views • 18 days ago

Robots need to be able to apply pressure and make contact with objects as needed in order to accomplish their tasks. From compliance to working safely around humans to whole-body manipulation of heavy objects, combining force and position control can dramatically expand the capabilities of robots. This is especially true for legged robots, which have so much ability to exert forces on the world around them. But how do we train robots which can do this? Baoxiong Jia tells us more in our discussion of his team’s recent, Best Paper Award winning work on learning a unified policy for position and force control, called UniFP. To learn more, watch Episode #49 of RoboPapers, hosted by Michael Cho - Rbt/Acc and Chris Paxton.

Robots need to be able to apply pressure and make contact with objects as needed in order to accomplish their tasks. From compliance to working safely around humans to whole-body manipulation of heavy objects, combining force and position control can dramatically expand the capabilities of robots. This is especially true for legged robots, which have so much ability to exert forces on the world around them. But how do we train robots which can do this? Baoxiong Jia tells us more in our discussion of his team’s recent, Best Paper Award winning work on learning a unified policy for position and force control, called UniFP. To learn more, watch Episode #49 of RoboPapers, hosted by Michael Cho - Rbt/Acc and Chris Paxton.

RoboPapers

44,803 views • 7 months ago

1/🧠Humans are the best robot data source — but video alone misses one thing: force. 2/🙁Tactile gloves capture force — but they're costly and block the real touch manipulation depends on. 3/💪Maybe the future of touch lives on your wrist: surface EMG reads the muscles that cause force — tactile sensing without ever touching a tactile sensor. 4/🔥Want a fully open-source framework — hardware + software — to train your own force-aware learn-from-human-data robot policy? 🚀We introduce ForceBand: Learning Forceful Manipulation with sEMG -- bring force into human videos with sEMG, for force-aware manipulation ⬇️ ✦ Zero-Shot Human-to-Robot Transfer ✦ Force Beyond Vision ✦ Free-Hand Force Sensing ✦ Collect by Anyone, Anytime, Anywhere ✦ Deploy on Any Robot, Any Camera, Any Environment ✦ Open-Source & Low-Cost & Easy-to-Implement Let's squeeze every bit of signal out of human data, and let robots feel the force! 🌐 Website: 📄 Paper: 💻 Code: 🎥 Video: 🧵 1/n

1/🧠Humans are the best robot data source — but video alone misses one thing: force. 2/🙁Tactile gloves capture force — but they're costly and block the real touch manipulation depends on. 3/💪Maybe the future of touch lives on your wrist: surface EMG reads the muscles that cause force — tactile sensing without ever touching a tactile sensor. 4/🔥Want a fully open-source framework — hardware + software — to train your own force-aware learn-from-human-data robot policy? 🚀We introduce ForceBand: Learning Forceful Manipulation with sEMG -- bring force into human videos with sEMG, for force-aware manipulation ⬇️ ✦ Zero-Shot Human-to-Robot Transfer ✦ Force Beyond Vision ✦ Free-Hand Force Sensing ✦ Collect by Anyone, Anytime, Anywhere ✦ Deploy on Any Robot, Any Camera, Any Environment ✦ Open-Source & Low-Cost & Easy-to-Implement Let's squeeze every bit of signal out of human data, and let robots feel the force! 🌐 Website: 📄 Paper: 💻 Code: 🎥 Video: 🧵 1/n

Zhi (Leo) Wang

50,156 views • 28 days ago

We're excited to share a new suite of AI agents that perform business research! 🤖 One of the most popular uses on Lutra is to perform web research on a large database or spreadsheet of accounts, leads, or potential hires. Having deep insights about them enables you to better qualify, prioritize, and even customize your approach. 👍 We are sharing a suite of data-driven playbooks that enables you to quickly understand what makes a business thrive - its team growth, tech stack, or social media engagements - just by plugging in a company name. These playbooks are designed to be easy to use: they come pre-built and you only need to provide a company name, or a spreadsheet of companies. They can surface the insights you need to guide strategy and win. 🚀 Like, repost, and find the link to the Playbooks in the thread! 🧵

We're excited to share a new suite of AI agents that perform business research! 🤖 One of the most popular uses on Lutra is to perform web research on a large database or spreadsheet of accounts, leads, or potential hires. Having deep insights about them enables you to better qualify, prioritize, and even customize your approach. 👍 We are sharing a suite of data-driven playbooks that enables you to quickly understand what makes a business thrive - its team growth, tech stack, or social media engagements - just by plugging in a company name. These playbooks are designed to be easy to use: they come pre-built and you only need to provide a company name, or a spreadsheet of companies. They can surface the insights you need to guide strategy and win. 🚀 Like, repost, and find the link to the Playbooks in the thread! 🧵

Jiquan Ngiam

23,148 views • 1 year ago

Can LLMs invent better ways to train LLMs? At Sakana AI, we’re pioneering AI-driven methods to automate AI research and discovery. We’re excited to release DiscoPOP: a new SOTA preference optimization algorithm that was discovered and written by an LLM! Our method leverages LLMs to propose and implement new preference optimization algorithms. We then train models with those algorithms and evaluate their performance, providing feedback to the LLM. By repeating this process for multiple generations in an evolutionary loop, the LLM discovers many highly-performant and novel preference optimization objectives! Paper: GitHub: Model: We proudly collaborated with the University of Oxford (Foerster Lab for AI Research (now part of BOLD)) and Cambridge University (Mihaela van der Schaar) on this groundbreaking project. Looking ahead, we envision a future where AI-driven research reduces the need for extensive human intervention and computational resources. This will accelerate scientific discoveries and innovation, pushing the boundaries of what AI can achieve.

Can LLMs invent better ways to train LLMs? At Sakana AI, we’re pioneering AI-driven methods to automate AI research and discovery. We’re excited to release DiscoPOP: a new SOTA preference optimization algorithm that was discovered and written by an LLM! Our method leverages LLMs to propose and implement new preference optimization algorithms. We then train models with those algorithms and evaluate their performance, providing feedback to the LLM. By repeating this process for multiple generations in an evolutionary loop, the LLM discovers many highly-performant and novel preference optimization objectives! Paper: GitHub: Model: We proudly collaborated with the University of Oxford (Foerster Lab for AI Research (now part of BOLD)) and Cambridge University (Mihaela van der Schaar) on this groundbreaking project. Looking ahead, we envision a future where AI-driven research reduces the need for extensive human intervention and computational resources. This will accelerate scientific discoveries and innovation, pushing the boundaries of what AI can achieve.

Sakana AI

555,922 views • 2 years ago

Brandon Johnson continues to push the false narrative that crime is down, expecting us to believe it. But the facts and everyday experiences of Chicagoans tell a different story. No amount of pandering, whether to the public or specific communities, will cover up failed policies. Governor JB Pritzker and Mayor of Chicago Brandon Johnson can try to spin the data, but we see through it. Their policies prioritize criminals and Illegal Aliens over law-abiding Americans. That’s unacceptable. We’re ready for change. We’re going to flip Chicago and Illinois red, with policies that prioritize safety, reward hard work, and put citizens first. Chicago Flips Red Fox News Donald J. Trump

Brandon Johnson continues to push the false narrative that crime is down, expecting us to believe it. But the facts and everyday experiences of Chicagoans tell a different story. No amount of pandering, whether to the public or specific communities, will cover up failed policies. Governor JB Pritzker and Mayor of Chicago Brandon Johnson can try to spin the data, but we see through it. Their policies prioritize criminals and Illegal Aliens over law-abiding Americans. That’s unacceptable. We’re ready for change. We’re going to flip Chicago and Illinois red, with policies that prioritize safety, reward hard work, and put citizens first. Chicago Flips Red Fox News Donald J. Trump

Danielle Carter

23,889 views • 1 year ago

ELON: NEURALINK’S TELEPATHY ENABLES YOU TO CONTROL ALMOST ANYTHING BY THINKING “Our first product, we call it Telepathy, enables the person with a Neuralink implant to control their phone or computer just by thinking. Once you can control your phone or computer, you can essentially control almost anything, literally just by thinking. There's no eye tracking or anything. It is purely your thoughts. This is really quite a profound device that can help a lot of people who have lost the connection between their brain and body. Imagine if people like Stephen Hawking could communicate at the same speed as someone who still had the connection to their brain and body. It's really something that can help millions of people around the world.” Neuralink update, July 10, 2024

ELON: NEURALINK’S TELEPATHY ENABLES YOU TO CONTROL ALMOST ANYTHING BY THINKING “Our first product, we call it Telepathy, enables the person with a Neuralink implant to control their phone or computer just by thinking. Once you can control your phone or computer, you can essentially control almost anything, literally just by thinking. There's no eye tracking or anything. It is purely your thoughts. This is really quite a profound device that can help a lot of people who have lost the connection between their brain and body. Imagine if people like Stephen Hawking could communicate at the same speed as someone who still had the connection to their brain and body. It's really something that can help millions of people around the world.” Neuralink update, July 10, 2024

Mario Nawfal

230,350 views • 1 year ago

🚨“The Vatican is what started Pizzagate… I had personal experience with a company that makes what they say are ‘pizza ovens’ and they are pizza ovens, but they’re actually altars to Baal & Moloch for the burning of children… I know of more than one local & international pizza chain that you can call with a code in order to order a child…” - SRA Survivor & Ex-Government Level Mind Control Programmer, Kadaya🚨 🎙️FULL EPISODE: “Kadaya - Dangerous Confessions of a Government-Level Mind Control Programmer & MONARCH Slave” 🔗

🚨“The Vatican is what started Pizzagate… I had personal experience with a company that makes what they say are ‘pizza ovens’ and they are pizza ovens, but they’re actually altars to Baal & Moloch for the burning of children… I know of more than one local & international pizza chain that you can call with a code in order to order a child…” - SRA Survivor & Ex-Government Level Mind Control Programmer, Kadaya🚨 🎙️FULL EPISODE: “Kadaya - Dangerous Confessions of a Government-Level Mind Control Programmer & MONARCH Slave” 🔗

E M M A ✨ K A T H E R I N E

216,731 views • 1 year ago

DisCo: Disentangled Control for Referring Human Dance Generation in Real World paper page: Generative AI has made significant strides in computer vision, particularly in image/video synthesis conditioned on text descriptions. Despite the advancements, it remains challenging especially in the generation of human-centric content such as dance synthesis. Existing dance synthesis methods struggle with the gap between synthesized content and real-world dance scenarios. In this paper, we define a new problem setting: Referring Human Dance Generation, which focuses on real-world dance scenarios with three important properties: (i) Faithfulness: the synthesis should retain the appearance of both human subject foreground and background from the reference image, and precisely follow the target pose; (ii) Generalizability: the model should generalize to unseen human subjects, backgrounds, and poses; (iii) Compositionality: it should allow for composition of seen/unseen subjects, backgrounds, and poses from different sources. To address these challenges, we introduce a novel approach, DISCO, which includes a novel model architecture with disentangled control to improve the faithfulness and compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans. Extensive qualitative and quantitative results demonstrate that DISCO can generate high-quality human dance images and videos with diverse appearances and flexible motions.

DisCo: Disentangled Control for Referring Human Dance Generation in Real World paper page: Generative AI has made significant strides in computer vision, particularly in image/video synthesis conditioned on text descriptions. Despite the advancements, it remains challenging especially in the generation of human-centric content such as dance synthesis. Existing dance synthesis methods struggle with the gap between synthesized content and real-world dance scenarios. In this paper, we define a new problem setting: Referring Human Dance Generation, which focuses on real-world dance scenarios with three important properties: (i) Faithfulness: the synthesis should retain the appearance of both human subject foreground and background from the reference image, and precisely follow the target pose; (ii) Generalizability: the model should generalize to unseen human subjects, backgrounds, and poses; (iii) Compositionality: it should allow for composition of seen/unseen subjects, backgrounds, and poses from different sources. To address these challenges, we introduce a novel approach, DISCO, which includes a novel model architecture with disentangled control to improve the faithfulness and compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans. Extensive qualitative and quantitative results demonstrate that DISCO can generate high-quality human dance images and videos with diverse appearances and flexible motions.

AK

161,453 views • 3 years ago

$*New Paper on AI & Democracy* Imagine two approaches to democracy. The one we have today, where citizens choose a professional politician to represent them and others. Or an augmented form of democracy, where each citizen controls a personalized AI that helps them participate in thousands of nuanced decisions. This second approach is the idea of Augmented Democracy I introduced six years ago at TED. In our latest paper we explore a simplified version of Augmented Democracy by combining off-the-shelf LLMs, such as ChatGPT, with data collected using a collaborative government program builder. This was an online game where people build a personalized government program using proposals extracted from the programs of the candidates of the 2022 presidential election in Brazil. So how accurate are these augmented forms of democracy? Imagine a user who gave us 40 answers. We can use the first 20 to fine-tune a model that we can test using the 20 answers the model didn’t see. We can then compare the accuracy of these predictions with the ones obtained by a “bundle” rule, which assumes that users that self-reported to be from the left or right always chose the proposals from the candidate that shares their political identity. This showed us that LLMs were more accurate at predicting policy preferences than the bundle rule, meaning that the preferences captured in the participation data were more nuanced than a left-right axis, and that the LLMs can capture some of that nuance. Also, the LLMs can choose among policies coming from the same candidate, which is something that we cannot do using a bundle rule. But can these LLMs help us complete the aggregate preferences of the population? Direct or unbundled forms of participation can result in incomplete data when people answer only a fraction of all questions. In our paper, we simulate this incompleteness by sampling the full dataset. We ask how close we can get to the full dataset by using a random sample, or a random sample augmented by predictions made by these LLMs. Overall, we find that LLM-augmented data gets much closer to the full dataset than a pure random sample. These results do not mean that augmented democracy technology is ready, but they means we are in a much better place to continue exploring this idea than six years ago. This paper was a collaborative effort with Jairo Gudino, PhD student at CCL at the University of Toulouse Capitole and Umberto Grandi from IRIT also at the University of Toulouse Capitole. We hope you find these results insightful!$

New Paper on AI & Democracy Imagine two approaches to democracy. The one we have today, where citizens choose a professional politician to represent them and others. Or an augmented form of democracy, where each citizen controls a personalized AI that helps them participate in thousands of nuanced decisions. This second approach is the idea of Augmented Democracy I introduced six years ago at TED. In our latest paper we explore a simplified version of Augmented Democracy by combining off-the-shelf LLMs, such as ChatGPT, with data collected using a collaborative government program builder. This was an online game where people build a personalized government program using proposals extracted from the programs of the candidates of the 2022 presidential election in Brazil. So how accurate are these augmented forms of democracy? Imagine a user who gave us 40 answers. We can use the first 20 to fine-tune a model that we can test using the 20 answers the model didn’t see. We can then compare the accuracy of these predictions with the ones obtained by a “bundle” rule, which assumes that users that self-reported to be from the left or right always chose the proposals from the candidate that shares their political identity. This showed us that LLMs were more accurate at predicting policy preferences than the bundle rule, meaning that the preferences captured in the participation data were more nuanced than a left-right axis, and that the LLMs can capture some of that nuance. Also, the LLMs can choose among policies coming from the same candidate, which is something that we cannot do using a bundle rule. But can these LLMs help us complete the aggregate preferences of the population? Direct or unbundled forms of participation can result in incomplete data when people answer only a fraction of all questions. In our paper, we simulate this incompleteness by sampling the full dataset. We ask how close we can get to the full dataset by using a random sample, or a random sample augmented by predictions made by these LLMs. Overall, we find that LLM-augmented data gets much closer to the full dataset than a pure random sample. These results do not mean that augmented democracy technology is ready, but they means we are in a much better place to continue exploring this idea than six years ago. This paper was a collaborative effort with Jairo Gudino, PhD student at CCL at the University of Toulouse Capitole and Umberto Grandi from IRIT also at the University of Toulouse Capitole. We hope you find these results insightful!

César A. Hidalgo

26,915 views • 1 year ago

"#Asseb has always been #Eritrean!" declares Dawit Woldegiorgis, a high level leader during the Dergue Era and was assigned by Mengistu to be #Eritrea's Governor during the heyday of the brutal regime makes it clear in a recent interview that #Ethiopia can never own a Sea Coast legally unless it uses force or #Eritreans willingly federate themselves with #Ethiopia. And we all know that will never happen.... #NoMoreWar

"#Asseb has always been #Eritrean!" declares Dawit Woldegiorgis, a high level leader during the Dergue Era and was assigned by Mengistu to be #Eritrea's Governor during the heyday of the brutal regime makes it clear in a recent interview that #Ethiopia can never own a Sea Coast legally unless it uses force or #Eritreans willingly federate themselves with #Ethiopia. And we all know that will never happen.... #NoMoreWar

Sirak Bahlbi

19,908 views • 9 months ago

VideoRF: Rendering Dynamic Radiance Fields as 2D Feature Video Streams paper page: Neural Radiance Fields (NeRFs) excel in photorealistically rendering static scenes. However, rendering dynamic, long-duration radiance fields on ubiquitous devices remains challenging, due to data storage and computational constraints. In this paper, we introduce VideoRF, the first approach to enable real-time streaming and rendering of dynamic radiance fields on mobile platforms. At the core is a serialized 2D feature image stream representing the 4D radiance field all in one. We introduce a tailored training scheme directly applied to this 2D domain to impose the temporal and spatial redundancy of the feature image stream. By leveraging the redundancy, we show that the feature image stream can be efficiently compressed by 2D video codecs, which allows us to exploit video hardware accelerators to achieve real-time decoding. On the other hand, based on the feature image stream, we propose a novel rendering pipeline for VideoRF, which has specialized space mappings to query radiance properties efficiently. Paired with a deferred shading model, VideoRF has the capability of real-time rendering on mobile devices thanks to its efficiency. We have developed a real-time interactive player that enables online streaming and rendering of dynamic scenes, offering a seamless and immersive free-viewpoint experience across a range of devices, from desktops to mobile phones.

VideoRF: Rendering Dynamic Radiance Fields as 2D Feature Video Streams paper page: Neural Radiance Fields (NeRFs) excel in photorealistically rendering static scenes. However, rendering dynamic, long-duration radiance fields on ubiquitous devices remains challenging, due to data storage and computational constraints. In this paper, we introduce VideoRF, the first approach to enable real-time streaming and rendering of dynamic radiance fields on mobile platforms. At the core is a serialized 2D feature image stream representing the 4D radiance field all in one. We introduce a tailored training scheme directly applied to this 2D domain to impose the temporal and spatial redundancy of the feature image stream. By leveraging the redundancy, we show that the feature image stream can be efficiently compressed by 2D video codecs, which allows us to exploit video hardware accelerators to achieve real-time decoding. On the other hand, based on the feature image stream, we propose a novel rendering pipeline for VideoRF, which has specialized space mappings to query radiance properties efficiently. Paired with a deferred shading model, VideoRF has the capability of real-time rendering on mobile devices thanks to its efficiency. We have developed a real-time interactive player that enables online streaming and rendering of dynamic scenes, offering a seamless and immersive free-viewpoint experience across a range of devices, from desktops to mobile phones.

AK

38,686 views • 2 years ago

Everything Elon said about Optimus at the All-In Summit today: • We’re finalizing the design of Optimus v3. That release is going to be a very remarkable robot. It will have manual dexterity comparable to a human, meaning a very complex hand, an AI mind that can navigate and comprehend reality, and will be made in very high volume. • Other robotics companies are missing those three very hard things. • I spend more mental cycles on Optimus than any other single thing. Solving real-world AI, all of the electrical-mechanical issues, the supply chain, and production challenges. • There is no supply chain for humanoid robots, so it has to be created from scratch, which requires a lot of vertical integration. None of the actuators in Optimus are available from an existing supply chain. • I think if successful, Optimus would be the biggest product ever. • The marginal cost of production, once we hit a million units per year, will probably be around $20,000. It depends on how much we spend on the AI chip in the robot, and we’ll need to achieve a lot of efficiencies in the actuators—26 actuators per arm (26 motors, gearboxes, and power electronics). The AI chip might cost $5,000 or $6,000, maybe more. At 1 million units a year, production cost will be $20,000, maybe $25,000. Price will be a function of demand. • Human hands have evolved to be incredibly sophisticated machines. Hands are a very first instrument. You can swing a baseball bat, thread a needle, play a piano or violin, and assemble a car. Hands are incredibly versatile instruments. Most of the muscles of the hands are actually in the forearm, and the hand is almost like a puppet. Human tendon evolution is incredibly good. The human hand has 27 or 28 degrees of freedom, depending on how you count it; it’s amazing. • In order to create a robot that can be a generalized humanoid, you must solve the “hands problem.” • Even though there are 10,000 to 20,000 electric motors out there, we couldn’t buy the actuators for any amount of money. We had to design every electric motor, gearbox, and controlling electronics from scratch, from first principles of physics. • Optimus is harder than developing any previous Tesla product, but not harder than Starship. • Right now, we’re struggling with the final design of the hardware, primarily the hand. The hands and forearm are the majority of the engineering difficulty of the entire robot. • If you want to do all the things that a human can do, it turns out you need a humanoid robot. If you want to do a subset, that’s much easier. Humans evolved to the shape and capability that we have for a good reason. There is value to having four fingers and a thumb; even the pinky is quite useful. Toes are much more of a question mark. • The AI5 inference chip will be 40 times better than AI4 by some measures. We know the limiting factors of the chip because the AI software and hardware teams work so closely. Effectively, the Tesla AI hardware and software teams are co-designing the chip. • The Softmax function on AI4 takes 40 steps in emulation mode, which will take only a few steps in AI5 natively. AI5 will easily handle mixed precision. • In terms of nominal raw compute, the AI5 inference chip has 8 times more compute, 9 times more memory, and 5 times more memory bandwidth compared to AI4. Because we’re addressing some core limitations and optimizations at the silicon level, we’re able to realize 40x improvements.

Everything Elon said about Optimus at the All-In Summit today: • We’re finalizing the design of Optimus v3. That release is going to be a very remarkable robot. It will have manual dexterity comparable to a human, meaning a very complex hand, an AI mind that can navigate and comprehend reality, and will be made in very high volume. • Other robotics companies are missing those three very hard things. • I spend more mental cycles on Optimus than any other single thing. Solving real-world AI, all of the electrical-mechanical issues, the supply chain, and production challenges. • There is no supply chain for humanoid robots, so it has to be created from scratch, which requires a lot of vertical integration. None of the actuators in Optimus are available from an existing supply chain. • I think if successful, Optimus would be the biggest product ever. • The marginal cost of production, once we hit a million units per year, will probably be around $20,000. It depends on how much we spend on the AI chip in the robot, and we’ll need to achieve a lot of efficiencies in the actuators—26 actuators per arm (26 motors, gearboxes, and power electronics). The AI chip might cost $5,000 or $6,000, maybe more. At 1 million units a year, production cost will be $20,000, maybe $25,000. Price will be a function of demand. • Human hands have evolved to be incredibly sophisticated machines. Hands are a very first instrument. You can swing a baseball bat, thread a needle, play a piano or violin, and assemble a car. Hands are incredibly versatile instruments. Most of the muscles of the hands are actually in the forearm, and the hand is almost like a puppet. Human tendon evolution is incredibly good. The human hand has 27 or 28 degrees of freedom, depending on how you count it; it’s amazing. • In order to create a robot that can be a generalized humanoid, you must solve the “hands problem.” • Even though there are 10,000 to 20,000 electric motors out there, we couldn’t buy the actuators for any amount of money. We had to design every electric motor, gearbox, and controlling electronics from scratch, from first principles of physics. • Optimus is harder than developing any previous Tesla product, but not harder than Starship. • Right now, we’re struggling with the final design of the hardware, primarily the hand. The hands and forearm are the majority of the engineering difficulty of the entire robot. • If you want to do all the things that a human can do, it turns out you need a humanoid robot. If you want to do a subset, that’s much easier. Humans evolved to the shape and capability that we have for a good reason. There is value to having four fingers and a thumb; even the pinky is quite useful. Toes are much more of a question mark. • The AI5 inference chip will be 40 times better than AI4 by some measures. We know the limiting factors of the chip because the AI software and hardware teams work so closely. Effectively, the Tesla AI hardware and software teams are co-designing the chip. • The Softmax function on AI4 takes 40 steps in emulation mode, which will take only a few steps in AI5 natively. AI5 will easily handle mixed precision. • In terms of nominal raw compute, the AI5 inference chip has 8 times more compute, 9 times more memory, and 5 times more memory bandwidth compared to AI4. Because we’re addressing some core limitations and optimizations at the silicon level, we’re able to realize 40x improvements.

The Humanoid Hub

238,630 views • 10 months ago