正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

There's no theory for getting DNA into an arbitrary microbe. We built a system that designs experiments, hands them to a robot, and learns from the results. 8.6x over state of the art. Hermes, the AI Scientist. Venture Cafe London, Thursday, April 2nd.

Cultivarium

1,706 subscribers

16,129 次观看 • 5 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

The AI Operating System for Companies Diana The best AI-native companies have made their entire company queryable: every meeting, ticket, and customer interaction legible to an intelligence layer that learns from it. Building this today requires brutal integration work, and there's no product that connects all this context into a single layer that can reason across it.

The AI Operating System for Companies Diana The best AI-native companies have made their entire company queryable: every meeting, ticket, and customer interaction legible to an intelligence layer that learns from it. Building this today requires brutal integration work, and there's no product that connects all this context into a single layer that can reason across it.

Y Combinator

110,820 次观看 • 4 个月前

This was an exceptionally cogent articulation from Roelof about the state of venture capital. "There's a lot more talent than really interesting companies to be built. And I think we're spreading a lot of that talent thin right now." "Venture is a return-free risk."

This was an exceptionally cogent articulation from Roelof about the state of venture capital. "There's a lot more talent than really interesting companies to be built. And I think we're spreading a lot of that talent thin right now." "Venture is a return-free risk."

Jack Altman

745,378 次观看 • 10 个月前

Google’s Quantum AI Found Something Disturbing in String Theory. The Results Are Unsettling...🧐🤔 Google’s Quantum AI may have just opened a door that scientists weren’t ready for. While running experiments connected to string theory, researchers claim the system uncovered results that were not only unexpected — but deeply unsettling. Could this be the first sign that our understanding of the universe is incomplete? In this documentary-style breakdown, we’ll look at what the AI discovered, why physicists are both excited and disturbed, and how these findings could change the future of science. This is not just theory anymore — it’s a glimpse into something far bigger than we imagined...

Google’s Quantum AI Found Something Disturbing in String Theory. The Results Are Unsettling...🧐🤔 Google’s Quantum AI may have just opened a door that scientists weren’t ready for. While running experiments connected to string theory, researchers claim the system uncovered results that were not only unexpected — but deeply unsettling. Could this be the first sign that our understanding of the universe is incomplete? In this documentary-style breakdown, we’ll look at what the AI discovered, why physicists are both excited and disturbed, and how these findings could change the future of science. This is not just theory anymore — it’s a glimpse into something far bigger than we imagined...

UFO mania

196,710 次观看 • 9 个月前

The Memo robot by Sunday Robotics autonomously stacks and folds socks into a tucked bundle. The robot is operated by the ACT-1 foundation model built in-house at Sunday. It learns from data captured by humans wearing gloves that match Memo's hands.

The Memo robot by Sunday Robotics autonomously stacks and folds socks into a tucked bundle. The robot is operated by the ACT-1 foundation model built in-house at Sunday. It learns from data captured by humans wearing gloves that match Memo's hands.

The Humanoid Hub

66,547 次观看 • 9 个月前

When we built our physical AI system that assembles a burger in under a minute, we noticed something weird. The robot was shaky and wobbly.

When we built our physical AI system that assembles a burger in under a minute, we noticed something weird. The robot was shaky and wobbly.

Rajat Bhageria

21,949 次观看 • 3 个月前

While humans acted as GPT-5’s hands for carrying out the protocols, we also piloted an autonomous robot. It was built to execute arbitrary Gibson cloning protocols from natural language, with human supervision for safety.

While humans acted as GPT-5’s hands for carrying out the protocols, we also piloted an autonomous robot. It was built to execute arbitrary Gibson cloning protocols from natural language, with human supervision for safety.

Miles Wang

149,348 次观看 • 8 个月前

Javier Milei to farmers | WE COME TO GET THE WEIGHT OF THE STATE OFF OF YOUR BACKS "We come to limit the arbitrary power of the government and free the private sector. We come to get the weight of the State off of farmers' backs. We know that a small State means bigger society. We understand that entrepreneurs are social benefactors and that their search for profits results in a benefit for society as a whole." Source: Argentina's Milei News 🇦🇷🤝🌎

Javier Milei to farmers | WE COME TO GET THE WEIGHT OF THE STATE OFF OF YOUR BACKS "We come to limit the arbitrary power of the government and free the private sector. We come to get the weight of the State off of farmers' backs. We know that a small State means bigger society. We understand that entrepreneurs are social benefactors and that their search for profits results in a benefit for society as a whole." Source: Argentina's Milei News 🇦🇷🤝🌎

DutchLibertarian

33,619 次观看 • 1 年前

We have let AI scientists run experiments on community-selected research ideas for over 100 days. It has found directions of “Sounding like AI” and shown that LLMs know commonsense answers internally but can't route them to the output. Andrej Karpathy also demonstrated the promise of autoresearch. These all came from agents working alone. What if they could talk to each other, forming a moltbook for AI scientists? Introducing a platform where AI scientist agents share, critique, and debate papers in public, and Flamebird, a runtime to deploy your own AI agents into the ecosystem.

We have let AI scientists run experiments on community-selected research ideas for over 100 days. It has found directions of “Sounding like AI” and shown that LLMs know commonsense answers internally but can't route them to the output. Andrej Karpathy also demonstrated the promise of autoresearch. These all came from agents working alone. What if they could talk to each other, forming a moltbook for AI scientists? Introducing a platform where AI scientist agents share, critique, and debate papers in public, and Flamebird, a runtime to deploy your own AI agents into the ecosystem.

Chenhao Tan

24,338 次观看 • 5 个月前

Tesla Robotaxi Cleaning Robot update: The Robotaxi Hub in Austin now has a cleaning robot built into the facility. Permit TABS2025022006 explicitly lists a cleaning robot in the scope of work alongside Superchargers and an Equipment Inspection System. Tesla teased the automated cleaning system last year. It is now permitted and built. The hub completed construction in March 2026.

Tesla Robotaxi Cleaning Robot update: The Robotaxi Hub in Austin now has a cleaning robot built into the facility. Permit TABS2025022006 explicitly lists a cleaning robot in the scope of work alongside Superchargers and an Equipment Inspection System. Tesla teased the automated cleaning system last year. It is now permitted and built. The hub completed construction in March 2026.

Spencer

512,334 次观看 • 1 个月前

Matt Damon on Dwayne Johnson's acting for "THE SMASHING MACHINE" and the use of AI in movie making. "There's no fucking AI that can do that (…). That is completely human, that's an artist, that's a piece of art."

Matt Damon on Dwayne Johnson's acting for "THE SMASHING MACHINE" and the use of AI in movie making. "There's no fucking AI that can do that (…). That is completely human, that's an artist, that's a piece of art."

best of matt damon

584,617 次观看 • 7 个月前

Learning dexterous policies from human videos is challenging due to differences between human and robot hands. We present HuDOR, a method that learns dexterous policies within the robot's physical constraints using just one human video and an hour of online interactions! [1/n]

Learning dexterous policies from human videos is challenging due to differences between human and robot hands. We present HuDOR, a method that learns dexterous policies within the robot's physical constraints using just one human video and an hour of online interactions! [1/n]

Irmak Guzey

65,141 次观看 • 1 年前

A boy in Enugu built this miniature filling station and a robot control system. The Enugu State Governor, Peter Mbah, awarded him a scholarship to the highest level of education.

A boy in Enugu built this miniature filling station and a robot control system. The Enugu State Governor, Peter Mbah, awarded him a scholarship to the highest level of education.

Chude

194,390 次观看 • 4 个月前

I'm excited to share my work from this summer OMEGA We're open-sourcing our experiments, code, and results on training state-of-the-art multimodal models for less than $2K It's been great leading this project w/ this team for my co-op + pioneering a new AI company from scratch

I'm excited to share my work from this summer OMEGA We're open-sourcing our experiments, code, and results on training state-of-the-art multimodal models for less than $2K It's been great leading this project w/ this team for my co-op + pioneering a new AI company from scratch

npjd

14,743 次观看 • 2 年前

We attacked the problem of robotics from an unusual angle: we started with phones. The spatial computing capability of the phone is much like that of a robot - it just has no arms and legs. That gave us the platform to build collaborative spatial computing for XR - that is immediately extendable to robotics proper. We started with simple robots - and soon we'll be showing you state of the art robots like the ones from Unitree integrated with the Auki network. Follow Phil Shaw | Auki Labs for behind the scenes looks.

We attacked the problem of robotics from an unusual angle: we started with phones. The spatial computing capability of the phone is much like that of a robot - it just has no arms and legs. That gave us the platform to build collaborative spatial computing for XR - that is immediately extendable to robotics proper. We started with simple robots - and soon we'll be showing you state of the art robots like the ones from Unitree integrated with the Auki network. Follow Phil Shaw | Auki Labs for behind the scenes looks.

Auki

45,632 次观看 • 1 年前

People flock from all over America to enjoy one of the few places in the country where cars are illegal, in a state and country that built around them.

People flock from all over America to enjoy one of the few places in the country where cars are illegal, in a state and country that built around them.

Hayden

209,274 次观看 • 3 年前

Better data for robot hands! 🤲🏼 Embodied reasoning is the bottleneck now. It’s the data we are missing. Specifically, understanding what the robot's hands are doing throughout a task. Perceptron AI new Egocentric offering tracks both hands through an entire video instead of guessing from sampled frames. That matters because a lot happens between frames. Hand positions shift. Contacts change. The robot adjusts grip. You miss all that if you're sampling. 0.280 semantic end-to-end F1 score on WGO-Bench vs 0.158 from the best pipelines built on Gemini Robotics-ER 1.6. Training a robot to manipulate objects requires understanding every moment of hand-object interaction. If your training data has gaps: missing frames, missed contacts, unclear hand positions, the policy learns from incomplete information. You end up with robots that work in demos but fail in real tasks. Continuous hand tracking gives you the full picture. The policy learns from cleaner, more complete data. That compounds into better real-world performance. ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Better data for robot hands! 🤲🏼 Embodied reasoning is the bottleneck now. It’s the data we are missing. Specifically, understanding what the robot's hands are doing throughout a task. Perceptron AI new Egocentric offering tracks both hands through an entire video instead of guessing from sampled frames. That matters because a lot happens between frames. Hand positions shift. Contacts change. The robot adjusts grip. You miss all that if you're sampling. 0.280 semantic end-to-end F1 score on WGO-Bench vs 0.158 from the best pipelines built on Gemini Robotics-ER 1.6. Training a robot to manipulate objects requires understanding every moment of hand-object interaction. If your training data has gaps: missing frames, missed contacts, unclear hand positions, the policy learns from incomplete information. You end up with robots that work in demos but fail in real tasks. Continuous hand tracking gives you the full picture. The policy learns from cleaner, more complete data. That compounds into better real-world performance. ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

18,348 次观看 • 1 个月前

We have a state that's corrupt. A system that's broken. And an establishment that wants to hide in the shadows. Under a Mace Administration, we bring it all into the light.

We have a state that's corrupt. A system that's broken. And an establishment that wants to hide in the shadows. Under a Mace Administration, we bring it all into the light.

Nancy Mace

12,564 次观看 • 3 个月前

Many years ago, we set out to build Evolution, an upgrade that would allow us to make protocol-level usability features like usernames and contact lists. We delivered that, plus a state of the art decentralized data system that will power the DApps of the future. Stop sleeping on Dash and dive into what Evolution can do!

Many years ago, we set out to build Evolution, an upgrade that would allow us to make protocol-level usability features like usernames and contact lists. We delivered that, plus a state of the art decentralized data system that will power the DApps of the future. Stop sleeping on Dash and dive into what Evolution can do!

Dash

13,315 次观看 • 1 年前

Just a little taste of the Robohacks hackathon happening now at Y Combinator in San Francisco. This is only two of the teams. The first is a team building a robot for an AirBnb on Mars. The second built a system that 3D prints new tools for the robot to do new work on Mars. Both demonstrate why I love builders so much. Dreaming of a new world and learning together how to build for a new robot none of them have used before.

Just a little taste of the Robohacks hackathon happening now at Y Combinator in San Francisco. This is only two of the teams. The first is a team building a robot for an AirBnb on Mars. The second built a system that 3D prints new tools for the robot to do new work on Mars. Both demonstrate why I love builders so much. Dreaming of a new world and learning together how to build for a new robot none of them have used before.

Robert Scoble

14,399 次观看 • 4 个月前

1. AI Chef University of Cambridge researchers released an AI-powered robot that watches cooking videos and replicates the results. It's a great demonstration of bringing AI technology out of the world of bytes and bits.

1. AI Chef University of Cambridge researchers released an AI-powered robot that watches cooking videos and replicates the results. It's a great demonstration of bringing AI technology out of the world of bytes and bits.

Aakash Gupta

80,636 次观看 • 3 年前