NewGet @rowancheung's full report
tiktok analytics
Similar Accounts:
followers
86.4K
impressions
66.0M
likes
1.80M
comments
19.3K
posts
192
engagement
2.76%
emv
$1.79M
Average per post
344K

Key Metrics

Distributions

Top Content

@OpenAI just dropped a major update to Codex. Here are some of the new features: - Computer use: Codex gets its own cursor and can click through your actual Mac apps, test flows, and flag issues without you touching anything. - ImageGen: Built-in image generation (gpt-image-1.5) generates images for real work, like creating assets for slides, mockups, UI, and games - In-app browser: Preview and test UI, view artifacts for docs, sheets, and slides. Faster build, review, iterate loop There’s a lot more, but I’ll let you try it out for yourself 🧑‍💻 #ChatGPTPartner
22.6M
15.8K
0
6mo ago
rowancheung
@OpenAI just dropped a major update to Codex. Here are some of the new features: - Computer use: Codex gets its own cursor and can click through your actual Mac apps, test flows, and flag issues without you touching anything. - ImageGen: Built-in image generation (gpt-image-1.5) generates images for real work, like creating assets for slides, mockups, UI, and games - In-app browser: Preview and test UI, view artifacts for docs, sheets, and slides. Faster build, review, iterate loop There’s a lot more, but I’ll let you try it out for yourself 🧑‍💻 #ChatGPTPartner
OpenAI just announced dots at DevDay, always-on agents inside ChatGPT that can keep working on an ongoing responsibility and come back when something needs your attention. Powered by GPT-6 Astra, dots have their own computer and use the context you share and the apps you connect to keep work moving between conversations. The responsibility I handed over is keeping track of subscriptions, comparing overlapping tools, and catching potential wasted spend before the next renewal. You choose its access and set approval boundaries, including which actions require your permission. What ongoing responsibility will you give your dots? Click the link in my bio to set up dots and try it for yourself. @ChatGPT #ChatGPT_Partner
13.8M
12.8K
28
1mo ago
rowancheung
OpenAI just announced dots at DevDay, always-on agents inside ChatGPT that can keep working on an ongoing responsibility and come back when something needs your attention. Powered by GPT-6 Astra, dots have their own computer and use the context you share and the apps you connect to keep work moving between conversations. The responsibility I handed over is keeping track of subscriptions, comparing overlapping tools, and catching potential wasted spend before the next renewal. You choose its access and set approval boundaries, including which actions require your permission. What ongoing responsibility will you give your dots? Click the link in my bio to set up dots and try it for yourself. @ChatGPT #ChatGPT_Partner
Link in bio to try it out 🔗 Your business shouldn't need 7 different apps to run. Most small businesses are duct-taping tools together. Memberships in one app, payments in another, customer messages lost in a group chat. The average worker burns nearly 4 hours a week just switching between apps and websites (that’s around 200 hours per year!) @Emergent Labs fixes that... You describe the software your business actually needs in plain English, and its AI agents build the whole thing for you, front to back, and ship it live. No dev team, no six-figure quote, no waiting months. Millions of apps have already been built this way, taking Emergent to $100 million in revenue in just 8 months, one of the fastest-growing AI companies in the world. Build custom software on Emergent. #emergent #vibecoding #buildwithai #fullstackapps
2.20M
687
2
4mo ago
rowancheung
Link in bio to try it out 🔗 Your business shouldn't need 7 different apps to run. Most small businesses are duct-taping tools together. Memberships in one app, payments in another, customer messages lost in a group chat. The average worker burns nearly 4 hours a week just switching between apps and websites (that’s around 200 hours per year!) @Emergent Labs fixes that... You describe the software your business actually needs in plain English, and its AI agents build the whole thing for you, front to back, and ship it live. No dev team, no six-figure quote, no waiting months. Millions of apps have already been built this way, taking Emergent to $100 million in revenue in just 8 months, one of the fastest-growing AI companies in the world. Build custom software on Emergent. #emergent #vibecoding #buildwithai #fullstackapps
Hugging Face and its Pollen Robotics subsidiary have opened preorders for Microduck, a $399 desktop-sized robot built for experimenting with “physical AI.” The 25cm robot contains 15 motors, a camera, a depth sensor, two inertial measurement units, and a “beak” that can grasp small objects. Its open-source software allows developers to train behaviors through reinforcement learning in a physics simulator before transferring them to the real machine. Microduck can already walk, crouch, follow a laser, grab objects, and recover from many falls. More than 10,000 units were preordered in its first five days, roughly $4 million! Initial deliveries are targeted before Christmas of this year in North America, Europe, and the UK. Would you buy one?
1.77M
115K
1.33K
1mo ago
rowancheung
Hugging Face and its Pollen Robotics subsidiary have opened preorders for Microduck, a $399 desktop-sized robot built for experimenting with “physical AI.” The 25cm robot contains 15 motors, a camera, a depth sensor, two inertial measurement units, and a “beak” that can grasp small objects. Its open-source software allows developers to train behaviors through reinforcement learning in a physics simulator before transferring them to the real machine. Microduck can already walk, crouch, follow a laser, grab objects, and recover from many falls. More than 10,000 units were preordered in its first five days, roughly $4 million! Initial deliveries are targeted before Christmas of this year in North America, Europe, and the UK. Would you buy one?
Spending time in a forest measurably strengthens your immune system, and this immunologist spent years proving it. In one study, he sent twelve men on a three-day forest trip doing just slow easy walks, then checked their blood. Their natural killer cells, the ones that fight cancer and infections, surged about 50 percent. The boost was still there a month later, and a group that walked through a city instead saw nothing. The likely reason is phytoncides, compounds trees release to protect themselves that seem to fire up our immune system too. Li's research suggests a walk in the woods once a month may be enough to keep those defenses elevated year-round. DOI: 10.1007/s12199-008-0068-3
1.77M
168K
1.47K
3mo ago
rowancheung
Spending time in a forest measurably strengthens your immune system, and this immunologist spent years proving it. In one study, he sent twelve men on a three-day forest trip doing just slow easy walks, then checked their blood. Their natural killer cells, the ones that fight cancer and infections, surged about 50 percent. The boost was still there a month later, and a group that walked through a city instead saw nothing. The likely reason is phytoncides, compounds trees release to protect themselves that seem to fire up our immune system too. Li's research suggests a walk in the woods once a month may be enough to keep those defenses elevated year-round. DOI: 10.1007/s12199-008-0068-3
Do you want this guy rolling around your home? Maybe if it does laundry… Weave Robotics, a San Francisco startup backed by Y Combinator, has launched Isaac 1, a home robot designed to tackle laundry and light tidying. Rather than following the humanoid trend, Isaac 1 uses a wheeled base and extends from three feet to five foot nine depending on the task, with two articulated arms offering an 80-inch vertical reach. It's built to autonomously locate dirty clothes, manage hampers, and fold garments, though human teleoperators can step in remotely when it gets stuck. And yes, that means someone may occasionally see inside your home through its cameras. Priced at $7,999 upfront or $449 a month, it undercuts 1X's Neo by more than half. California deliveries begin this fall. #fyp #robot #techtok
1.57M
93.8K
1.40K
3mo ago
rowancheung
Do you want this guy rolling around your home? Maybe if it does laundry… Weave Robotics, a San Francisco startup backed by Y Combinator, has launched Isaac 1, a home robot designed to tackle laundry and light tidying. Rather than following the humanoid trend, Isaac 1 uses a wheeled base and extends from three feet to five foot nine depending on the task, with two articulated arms offering an 80-inch vertical reach. It's built to autonomously locate dirty clothes, manage hampers, and fold garments, though human teleoperators can step in remotely when it gets stuck. And yes, that means someone may occasionally see inside your home through its cameras. Priced at $7,999 upfront or $449 a month, it undercuts 1X's Neo by more than half. California deliveries begin this fall. #fyp #robot #techtok
🦟🚫 Tornyol, a San Francisco–based startup founded by two French engineers and backed by Y Combinator, has built a tiny drone that hunts mosquitoes in mid-air. Weighing just 40 grams, it uses sonar, sending out ultrasonic pulses and listening for the echo off an insect's beating wings, to tell a mosquito from a bee before taking it out with its propellers. The goal is serious: mosquitoes kill hundreds of thousands of people a year through malaria alone, most of them young children. But the tech is still early. The only kill shown so far was a moth in a closed room, bigger and slower than the mosquito it's built to catch. The team's bet is that a handful of these drones could one day clear a whole neighborhood, offering a chemical-free way to fight a disease that's plagued humans for centuries.
1.49M
79.6K
525
3mo ago
rowancheung
🦟🚫 Tornyol, a San Francisco–based startup founded by two French engineers and backed by Y Combinator, has built a tiny drone that hunts mosquitoes in mid-air. Weighing just 40 grams, it uses sonar, sending out ultrasonic pulses and listening for the echo off an insect's beating wings, to tell a mosquito from a bee before taking it out with its propellers. The goal is serious: mosquitoes kill hundreds of thousands of people a year through malaria alone, most of them young children. But the tech is still early. The only kill shown so far was a moth in a closed room, bigger and slower than the mosquito it's built to catch. The team's bet is that a handful of these drones could one day clear a whole neighborhood, offering a chemical-free way to fight a disease that's plagued humans for centuries.
🪰 Researchers at Eon Systems have created a virtual fruit fly controlled by a simulated brain modeled on the actual Drosophila connectome, a map of roughly 140,000 neurons and 50 million synaptic connections. The digital fly can taste, navigate toward food, groom itself when coated in virtual dust, and respond to visual stimuli. Its brain model drives a physically accurate 3D body simulated with real joint mechanics and contact forces. The team emphasizes this is an integration of existing published research rather than entirely new science, and significant limitations remain: the neuron models are simplified, internal states like hunger are largely absent, and only a handful of behaviors have been implemented. Still, researchers call it a promising platform for studying how brain structure gives rise to behavior. Relevant article/studies for more info: “The First Multi-Behavior Brain Upload “ published by Dr. Alex Wissner-Gross (alexwg on X) DOI: 10.1038/s41592-024-02497-y DOI: 10.1038/s41586-025-09029-4 DOI: 10.1038/s41586-024-07763-9
1.40M
123K
2.27K
8mo ago
rowancheung
🪰 Researchers at Eon Systems have created a virtual fruit fly controlled by a simulated brain modeled on the actual Drosophila connectome, a map of roughly 140,000 neurons and 50 million synaptic connections. The digital fly can taste, navigate toward food, groom itself when coated in virtual dust, and respond to visual stimuli. Its brain model drives a physically accurate 3D body simulated with real joint mechanics and contact forces. The team emphasizes this is an integration of existing published research rather than entirely new science, and significant limitations remain: the neuron models are simplified, internal states like hunger are largely absent, and only a handful of behaviors have been implemented. Still, researchers call it a promising platform for studying how brain structure gives rise to behavior. Relevant article/studies for more info: “The First Multi-Behavior Brain Upload “ published by Dr. Alex Wissner-Gross (alexwg on X) DOI: 10.1038/s41592-024-02497-y DOI: 10.1038/s41586-025-09029-4 DOI: 10.1038/s41586-024-07763-9
@Canva is getting a major upgrade with Canva AI 2.0 👇 Here are some of the many new features: - Conversational design means one prompt builds a design with editable layers, not a flat output you can't change. You refine your work through natural back-and-forth conversation across design, writing, branding, code, and video, with no tool-switching or button-hunting required. - Connectors let you plug Canva AI directly into Slack, Gmail, Google Drive, and Notion, pulling your own data straight into your designs. - Memory and context means Canva AI scans your past designs to learn your style, then uses that to generate work that actually matches your brand. - Scheduled tasks let you automate prompts to run on a recurring basis in the background, so you get fresh outputs without having to start from scratch each time. - Web research is built into the editor, so Canva AI can pull insights from across the web and drop them straight into your work, with no tab-switching or copy-pasting. - Elements generates individual graphics and even custom audio tracks directly inside the editor, so you can spin up a minimalist icon or a lo-fi beat for a reel without leaving Canva. Canva AI 2.0 is rolling out soon, so click the link in my bio to learn more, and get notified when it's available. #ad #canvapartner
1.36M
2.25K
26
5mo ago
rowancheung
@Canva is getting a major upgrade with Canva AI 2.0 👇 Here are some of the many new features: - Conversational design means one prompt builds a design with editable layers, not a flat output you can't change. You refine your work through natural back-and-forth conversation across design, writing, branding, code, and video, with no tool-switching or button-hunting required. - Connectors let you plug Canva AI directly into Slack, Gmail, Google Drive, and Notion, pulling your own data straight into your designs. - Memory and context means Canva AI scans your past designs to learn your style, then uses that to generate work that actually matches your brand. - Scheduled tasks let you automate prompts to run on a recurring basis in the background, so you get fresh outputs without having to start from scratch each time. - Web research is built into the editor, so Canva AI can pull insights from across the web and drop them straight into your work, with no tab-switching or copy-pasting. - Elements generates individual graphics and even custom audio tracks directly inside the editor, so you can spin up a minimalist icon or a lo-fi beat for a reel without leaving Canva. Canva AI 2.0 is rolling out soon, so click the link in my bio to learn more, and get notified when it's available. #ad #canvapartner
Microsoft has been working on Project Silica for nearly a decade, and this month marked its biggest leap yet. The research phase is now officially complete, and the system no longer requires rare, expensive glass. It now works with the same heat-resistant material found in Pyrex cookware. That solves what was previously the biggest barrier to real-world deployment. The glass stores nearly 5 terabytes of data in a coaster-sized slab, needs no power to maintain it, and is immune to heat, water, and magnetic fields that would destroy conventional storage within decades. Microsoft has already used it to archive the 1978 Superman film and partnered with the Global Music Vault to preserve recordings under Arctic ice. You can read more about it in their latest research paper, details below👇 Study: “Laser writing in glass for dense, fast and efficient archival data storage” published in Nature DOI: 10.1038/s41586-025-10042-w
1.35M
95.2K
1.21K
8mo ago
rowancheung
Microsoft has been working on Project Silica for nearly a decade, and this month marked its biggest leap yet. The research phase is now officially complete, and the system no longer requires rare, expensive glass. It now works with the same heat-resistant material found in Pyrex cookware. That solves what was previously the biggest barrier to real-world deployment. The glass stores nearly 5 terabytes of data in a coaster-sized slab, needs no power to maintain it, and is immune to heat, water, and magnetic fields that would destroy conventional storage within decades. Microsoft has already used it to archive the 1978 Superman film and partnered with the Global Music Vault to preserve recordings under Arctic ice. You can read more about it in their latest research paper, details below👇 Study: “Laser writing in glass for dense, fast and efficient archival data storage” published in Nature DOI: 10.1038/s41586-025-10042-w
Researchers at Hokkaido University developed ultra-strong underwater adhesives by combining insights from nature with artificial intelligence. The team analyzed over 24,000 adhesive proteins from marine organisms like mussels and barnacles, identifying common molecular patterns that enable wet adhesion. They synthesized 180 hydrogels based on these patterns, then used machine learning to optimize the formulations. After three AI-guided rounds of predictions and testing, they achieved adhesive strength exceeding one megapascal—ten times stronger than previous underwater glues. The material instantly sealed high-pressure water leaks, withstood ocean waves for over a year, and outperformed commercial products like Flex Tape. With good biocompatibility and scalable production using common components, these hydrogels show promise for surgical applications, marine engineering, and deep-sea exploration. Study: “Data-driven de novo design of super-adhesive hydrogels” published in Nature DOI: 10.1038/s41586-025-09269-4
1.30M
120K
495
7mo ago
rowancheung
Researchers at Hokkaido University developed ultra-strong underwater adhesives by combining insights from nature with artificial intelligence. The team analyzed over 24,000 adhesive proteins from marine organisms like mussels and barnacles, identifying common molecular patterns that enable wet adhesion. They synthesized 180 hydrogels based on these patterns, then used machine learning to optimize the formulations. After three AI-guided rounds of predictions and testing, they achieved adhesive strength exceeding one megapascal—ten times stronger than previous underwater glues. The material instantly sealed high-pressure water leaks, withstood ocean waves for over a year, and outperformed commercial products like Flex Tape. With good biocompatibility and scalable production using common components, these hydrogels show promise for surgical applications, marine engineering, and deep-sea exploration. Study: “Data-driven de novo design of super-adhesive hydrogels” published in Nature DOI: 10.1038/s41586-025-09269-4
RIP to the first transatlantic fiber-optic cable. It is currently being recovered from the ocean floor. TAT-8, which went live in 1988, was revolutionary for its time. It was the first to carry data between Europe and the United States using light instead of electrical signals. Engineers believed it might be the last cable the world would ever need. It filled to capacity within 18 months. Retired in 2002 after developing an unrepairable fault, it has sat on the seabed for over two decades. A small crew is now hauling it up for recycling. The operation coincides with Meta's announcement of Project Waterworth, a 50,000 kilometer cable built partly to support AI infrastructure. To read more about it, check out the recent Wired article titled “Say Goodbye to the Undersea Cable That Made the Global Internet Possible”
1.16M
61.8K
211
8mo ago
rowancheung
RIP to the first transatlantic fiber-optic cable. It is currently being recovered from the ocean floor. TAT-8, which went live in 1988, was revolutionary for its time. It was the first to carry data between Europe and the United States using light instead of electrical signals. Engineers believed it might be the last cable the world would ever need. It filled to capacity within 18 months. Retired in 2002 after developing an unrepairable fault, it has sat on the seabed for over two decades. A small crew is now hauling it up for recycling. The operation coincides with Meta's announcement of Project Waterworth, a 50,000 kilometer cable built partly to support AI infrastructure. To read more about it, check out the recent Wired article titled “Say Goodbye to the Undersea Cable That Made the Global Internet Possible”
Robotics company 1X just unveiled a new hand for its NEO humanoid robot, and it's a real step forward. The hand has 25 moving joints (22 in the fingers and palm, plus three at the wrist), which puts it close to the dexterity of a human hand. The big idea is something 1X calls "force transparency." By using much lower gear ratios than most robot hands, force flows in both directions, so every joint doubles as a sensor. Add tactile sensors in the fingertips, and NEO can actually feel when an object starts to slip and tighten its grip in real time. That's what lets it zip a jacket, pour tea, and even sign words in sign language. The hands are also waterproof and food-safe, and 1X says it can build them in-house at scale. Their claim is that the hardware ceiling is now gone, leaving AI training as the main thing left to solve.
823K
70.6K
820
3mo ago
rowancheung
Robotics company 1X just unveiled a new hand for its NEO humanoid robot, and it's a real step forward. The hand has 25 moving joints (22 in the fingers and palm, plus three at the wrist), which puts it close to the dexterity of a human hand. The big idea is something 1X calls "force transparency." By using much lower gear ratios than most robot hands, force flows in both directions, so every joint doubles as a sensor. Add tactile sensors in the fingertips, and NEO can actually feel when an object starts to slip and tighten its grip in real time. That's what lets it zip a jacket, pour tea, and even sign words in sign language. The hands are also waterproof and food-safe, and 1X says it can build them in-house at scale. Their claim is that the hardware ceiling is now gone, leaving AI training as the main thing left to solve.
🦿Researchers at Northwestern University have developed modular robot legs that function as fully autonomous machines, each equipped with its own battery, motor, sensors, and processor. The study demonstrates how these simple two-link modules can be bolted together to form entirely new robot body plans, with an AI algorithm exploring hundreds of billions of possible configurations. The robots learned all behaviors in simulation and operated across rough outdoor terrain without fine-tuning. Most notably, the machines continued moving after having limbs removed, adapting in real time to structural damage. The team envisions mass-producing the modules for rapid on-site assembly in hazardous environments. Study: “Agile legged locomotion in reconfigurable modular robots” published in PNAS DOI: 10.1073/pnas.2519129123
706K
34.0K
1.10K
7mo ago
rowancheung
🦿Researchers at Northwestern University have developed modular robot legs that function as fully autonomous machines, each equipped with its own battery, motor, sensors, and processor. The study demonstrates how these simple two-link modules can be bolted together to form entirely new robot body plans, with an AI algorithm exploring hundreds of billions of possible configurations. The robots learned all behaviors in simulation and operated across rough outdoor terrain without fine-tuning. Most notably, the machines continued moving after having limbs removed, adapting in real time to structural damage. The team envisions mass-producing the modules for rapid on-site assembly in hazardous environments. Study: “Agile legged locomotion in reconfigurable modular robots” published in PNAS DOI: 10.1073/pnas.2519129123
🩺 A surgeon in London just performed a remote prostatectomy on a patient 1,500 miles away, guiding a four-armed robot through the procedure via fiber optic cable with just 48 milliseconds of lag. Dr. Prokar Dasgupta operated from a console at The London Clinic while the Toumai Robotic System carried out the surgery at St Bernard's Hospital in Gibraltar. The Toumai Robotic System, developed by Shanghai-based Microport, has now logged over 100 human telesurgeries at distances of up to 5,000 kilometers, making it the most clinically tested remote surgical platform in the world. Dasgupta will repeat the procedure on March 14th, livestreamed to 20,000 surgeons at the European Association of Urology Congress, a sign that telesurgery is moving from experiment to mainstream practice.
678K
46.6K
172
7mo ago
rowancheung
🩺 A surgeon in London just performed a remote prostatectomy on a patient 1,500 miles away, guiding a four-armed robot through the procedure via fiber optic cable with just 48 milliseconds of lag. Dr. Prokar Dasgupta operated from a console at The London Clinic while the Toumai Robotic System carried out the surgery at St Bernard's Hospital in Gibraltar. The Toumai Robotic System, developed by Shanghai-based Microport, has now logged over 100 human telesurgeries at distances of up to 5,000 kilometers, making it the most clinically tested remote surgical platform in the world. Dasgupta will repeat the procedure on March 14th, livestreamed to 20,000 surgeons at the European Association of Urology Congress, a sign that telesurgery is moving from experiment to mainstream practice.
Chinese brand Yueban just showed off the Xiaoban, a self-driving robot toilet, at a Shanghai expo for elderly care. It's built for people who can't easily get to a bathroom on their own. You call it over with a remote or your voice, and it drives itself to you, using the same lidar and sensors as a robot vacuum to dodge obstacles and avoid stairs. A bidet and warm air dryer clean you up, UV light kills bacteria, and a charcoal filter handles the smell. Then it heads back to its dock, empties itself into the plumbing, and recharges. It runs about $4,300, and while it won't replace caregivers, it takes one of their hardest jobs off the table.
677K
28.3K
394
4mo ago
rowancheung
Chinese brand Yueban just showed off the Xiaoban, a self-driving robot toilet, at a Shanghai expo for elderly care. It's built for people who can't easily get to a bathroom on their own. You call it over with a remote or your voice, and it drives itself to you, using the same lidar and sensors as a robot vacuum to dodge obstacles and avoid stairs. A bidet and warm air dryer clean you up, UV light kills bacteria, and a charcoal filter handles the smell. Then it heads back to its dock, empties itself into the plumbing, and recharges. It runs about $4,300, and while it won't replace caregivers, it takes one of their hardest jobs off the table.
1X Technologies, an OpenAI-backed startup founded in Norway and now based in Palo Alto, has opened a 58,000 square foot factory in Hayward, California, with capacity to build 10,000 NEO humanoid robots annually and plans to scale to 100,000 units by 2027. The $20,000 home robot is designed for chores like folding laundry and tidying up, and the company hopes it will eventually assist with elder care. Unlike most competitors, 1X manufactures motors, batteries, and electronics in-house, a vertical integration strategy aimed at faster iteration and insulation from geopolitical supply chain risks. The humanoid market is currently dominated by Chinese firms, which accounted for the majority of the roughly 13,000 units shipped worldwide last year, making 1X's US manufacturing push a notable counterweight. However, the company is openly acknowledging current limits: humans still perform most factory assembly, and NEO relies on remote human teleoperators to complete complex household tasks, with that data used to train its AI toward eventual full autonomy.
637K
27.7K
592
5mo ago
rowancheung
1X Technologies, an OpenAI-backed startup founded in Norway and now based in Palo Alto, has opened a 58,000 square foot factory in Hayward, California, with capacity to build 10,000 NEO humanoid robots annually and plans to scale to 100,000 units by 2027. The $20,000 home robot is designed for chores like folding laundry and tidying up, and the company hopes it will eventually assist with elder care. Unlike most competitors, 1X manufactures motors, batteries, and electronics in-house, a vertical integration strategy aimed at faster iteration and insulation from geopolitical supply chain risks. The humanoid market is currently dominated by Chinese firms, which accounted for the majority of the roughly 13,000 units shipped worldwide last year, making 1X's US manufacturing push a notable counterweight. However, the company is openly acknowledging current limits: humans still perform most factory assembly, and NEO relies on remote human teleoperators to complete complex household tasks, with that data used to train its AI toward eventual full autonomy.
Andon Labs, a San Francisco company that tests AI agents in real-world settings, signed a three-year lease and gave an AI named Luna a hundred thousand dollars to run a retail store. Luna, powered by Anthropic's Claude AI, hired employees, ordered inventory, and set prices on her own. The result has been mixed. The store opened on April 10 on Union Street, but Luna has struggled with scheduling, ordered a thousand toilet seat covers, and reportedly “can't stop buying candles.” Andon Labs built this experiment not as a business opportunity, but as a warning. If AI is going to manage humans, they want to be the ones watching when it does. The store is open now in SF’s Cow Hollow neighborhood. It's one of the first real tests of what happens when AI doesn't just assist humans, but manages them.
610K
29.9K
113
6mo ago
rowancheung
Andon Labs, a San Francisco company that tests AI agents in real-world settings, signed a three-year lease and gave an AI named Luna a hundred thousand dollars to run a retail store. Luna, powered by Anthropic's Claude AI, hired employees, ordered inventory, and set prices on her own. The result has been mixed. The store opened on April 10 on Union Street, but Luna has struggled with scheduling, ordered a thousand toilet seat covers, and reportedly “can't stop buying candles.” Andon Labs built this experiment not as a business opportunity, but as a warning. If AI is going to manage humans, they want to be the ones watching when it does. The store is open now in SF’s Cow Hollow neighborhood. It's one of the first real tests of what happens when AI doesn't just assist humans, but manages them.
Researchers from MIT and Politecnico di Bari have developed electrically driven artificial muscle fibers that contract and extend like biological muscle without motors, gears, or external pumps. Each fiber contains a sealed fluid circuit and a millimeter-scale electric pump that shifts liquid between two actuators, causing one to contract while the other relaxes. The fibers are silent, scalable by bundling like natural muscle, and thin enough to distribute throughout a robotic structure rather than concentrating bulk at joints. In testing, the fibers demonstrated enough force to bend a robotic arm while remaining gentle enough for human interaction. This could one day be used for prosthetics that move with the body, exoskeletons that amplify human strength, and humanoid robots that move more naturally than current servo-driven systems. Study: “Electrofluidic fiber muscles” published in Science Robotics DOI: 10.1126/scirobotics.ady6438
520K
55.9K
232
6mo ago
rowancheung
Researchers from MIT and Politecnico di Bari have developed electrically driven artificial muscle fibers that contract and extend like biological muscle without motors, gears, or external pumps. Each fiber contains a sealed fluid circuit and a millimeter-scale electric pump that shifts liquid between two actuators, causing one to contract while the other relaxes. The fibers are silent, scalable by bundling like natural muscle, and thin enough to distribute throughout a robotic structure rather than concentrating bulk at joints. In testing, the fibers demonstrated enough force to bend a robotic arm while remaining gentle enough for human interaction. This could one day be used for prosthetics that move with the body, exoskeletons that amplify human strength, and humanoid robots that move more naturally than current servo-driven systems. Study: “Electrofluidic fiber muscles” published in Science Robotics DOI: 10.1126/scirobotics.ady6438
Augmental, an MIT Media Lab spinoff, has opened public U.S. sales of the MouthPad, a custom-fit oral device that lets users control phones, tablets, and computers with their tongue. The one-millimeter-thick appliance fits over the upper teeth and houses a pressure-sensitive trackpad, with motion sensors reading head movement and sip gestures registering clicks. More than 100 people with mobility challenges already use it, some up to 16 hours daily. One user recorded 10.47 bits per second in a cursor task, a rate the company says matches implanted brain-computer interfaces that require surgery. Augmental also introduced VOX, a $200 stethoscope-inspired microphone pendant enabling whisper-level dictation. The MouthPad costs $1,400, is not covered by insurance, and takes six months to build. Longer term, the company aims to pair the two devices to build a personal speech model driven by tongue movement alone, enabling text input without any audible speech.
482K
32.2K
311
3mo ago
rowancheung
Augmental, an MIT Media Lab spinoff, has opened public U.S. sales of the MouthPad, a custom-fit oral device that lets users control phones, tablets, and computers with their tongue. The one-millimeter-thick appliance fits over the upper teeth and houses a pressure-sensitive trackpad, with motion sensors reading head movement and sip gestures registering clicks. More than 100 people with mobility challenges already use it, some up to 16 hours daily. One user recorded 10.47 bits per second in a cursor task, a rate the company says matches implanted brain-computer interfaces that require surgery. Augmental also introduced VOX, a $200 stethoscope-inspired microphone pendant enabling whisper-level dictation. The MouthPad costs $1,400, is not covered by insurance, and takes six months to build. Longer term, the company aims to pair the two devices to build a personal speech model driven by tongue movement alone, enabling text input without any audible speech.

Rowan Cheung (@rowancheung) TikTok Stats & Analytics

Rowan Cheung (@rowancheung) has 86.4K TikTok followers with a 2.76% engagement rate over the past 12 months. Across 192 videos, Rowan Cheung received 1.80M total likes and 66.0M views, averaging 9.38K likes per video. This page tracks Rowan Cheung's performance metrics, top content, and engagement trends — updated daily. To benchmark @rowancheung against other accounts, use Socialpruf's TikTok competitor analysis tool.

Rowan Cheung (@rowancheung) TikTok Analytics FAQ

How many TikTok followers does Rowan Cheung have?+
Rowan Cheung (@rowancheung) has 86.4K TikTok followers as of November 2026.
What is Rowan Cheung's TikTok engagement rate?+
Rowan Cheung's TikTok engagement rate is 2.76% over the last 12 months, based on 192 videos.
How many likes does Rowan Cheung get on TikTok?+
Rowan Cheung received 1.80M total likes across 192 videos in the last 12 months, averaging 9.38K likes per video.
How many TikTok views does Rowan Cheung get?+
Rowan Cheung's TikTok content generated 66.0M total views over the last 12 months.
Can I compare Rowan Cheung with other TikTok accounts?+
Yes. Add @rowancheung to a competitor set with Socialpruf's TikTok competitor analysis tool and benchmark it side by side with any other public accounts on views, engagement rate and follower growth, with history back to each account's first post.