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。

Just submitted to arXiv.org: Amy's work on closed loop autonomous crystallisation, sample preparation, and powder X-ray diffraction. This is the most complex automated workflow that we've built so far, involving 3 separate robots & 13 steps Leverhulme Research Centre - Materials Design MIF ERC_ADAM

Andy Cooper

9,787 subscribers

39,425 次观看 • 2 年前 •via X (Twitter)

教育新闻政治科学技术

Anya Rossi• Live Now

Private livecam show

11 条评论

Andy Cooper 的头像

Andy Cooper2 年前

The workflow starts with crystal growth followed by sample preparation (2-step grinding), sample mounting, and PXRD data acquisition. It is orchestrated by our system architecture, #ARChemist, masterminded by @HatemFakhrulde1

Andy Cooper 的头像

Andy Cooper2 年前

A key step is grinding the crystals for better orientational averaging (and, indeed, to get them out of the sample vial ...)

Andy Cooper 的头像

Andy Cooper2 年前

To process the crystals, we use this versatile @ABBRobotics YuMi robot for sample handling

Andy Cooper 的头像

Andy Cooper2 年前

Very proud of @amylunt for building this in her PhD, especially given the huge slow-down during COVID. Amy, you're a superstar. The work was done with @sam_c and also involved @HatemFakhrulde1, @gabriellapizz, Louis, @TheWubberDuck, @nici_rankin, @robclow11 and Ben A. ⭐️⭐️⭐️

Andy Cooper 的头像

Andy Cooper2 年前

By the way - if anyone is interested, we are hiring a Lecturer in this general area of chemistry automation - advert to be posted soon (DM me if you're interested)

Andy Cooper 的头像

Andy Cooper2 年前

Credit also to @daftpunk for the 🎸

Andy Cooper 的头像

Andy Cooper2 年前

Btw, I have the @amazon receipt for the MP3 ...

Prof Ross Forgan 的头像

Prof Ross Forgan2 年前

@arxiv @amylunt @LC_Mater_Design @MIF_UoL @erc_adam This gets a “very cool!” and “how can they get robots to do that?” from my 5yo daughter. Have to say I’m in agreement!

Andy Cooper 的头像

Andy Cooper2 年前

@arxiv @amylunt @LC_Mater_Design @MIF_UoL @erc_adam Cheers Ross. The toys are in the post.

Pepe Marquez 的头像

Pepe Marquez2 年前

@MolecularXtal @arxiv @amylunt @LC_Mater_Design @MIF_UoL @erc_adam This is super cool Andy! How do you organize and store the huge amount of data that you guys generate? Did you code your own database?

Andy Cooper 的头像

Andy Cooper2 年前

@MolecularXtal @arxiv @amylunt @LC_Mater_Design @MIF_UoL @erc_adam Thanks! This is a recent implementation so we don't (yet) have a huge amount of data, but there is the potential to generate it - and yes, we're looking at databases also to compare with predicted structures with @graeme_day in our @erc_adam project

相关视频

US Secretary of State Marco Rubio: It is in my personal view that we've had probably the most productive and meaningful meeting so far in this entire process since we've been involved in it from the beginning. So I think the takeaway from it is I think this is a very, very meaningful, I would say probably best meeting and day we've had so far in this entire process going back to when we first came into office in January. We have a very good work product that was already built on a foundation of input from all the relative parties involved here, and we were able to go through some of those items now point by point, and I think we've made good progress. And then obviously there's the Russian side of the equation, but again, we think we have some pretty substantial insights over the last nine months into some of the things that are really important to them.

US Secretary of State Marco Rubio: It is in my personal view that we've had probably the most productive and meaningful meeting so far in this entire process since we've been involved in it from the beginning. So I think the takeaway from it is I think this is a very, very meaningful, I would say probably best meeting and day we've had so far in this entire process going back to when we first came into office in January. We have a very good work product that was already built on a foundation of input from all the relative parties involved here, and we were able to go through some of those items now point by point, and I think we've made good progress. And then obviously there's the Russian side of the equation, but again, we think we have some pretty substantial insights over the last nine months into some of the things that are really important to them.

Anton Gerashchenko

281,149 次观看 • 7 个月前

ELON: AGI IS LIKELY TO EMERGE FROM OUR AUTONOMOUS CARS AND HUMANOIDS “We're not actually specifically focused on AGI. I'm simply saying that AGI seems likely to be an emergent property of what we're doing, because we're creating all these autonomous cars and autonomous humanoids that are actually a truly gigantic data stream that's coming in and being processed. It's by far the most amount of real-world data, and data you can't get by just searching the Internet, because you have to be out there in the world and interacting with people and interacting with the roads. Earth is a big place, and reality is messy and complicated. If you've got tens or hundreds of millions of autonomous vehicles and maybe even a comparable number of humanoids, maybe more than that, on the humanoid front, well, that's just the most amount of data. And if that video is being processed, it just seems likely that the cars will definitely get way better than human drivers, and the humanoid robots will become increasingly indistinguishable from humans, perhaps. And so then, like you said, you have this emergent property of AGI.” Source: Tesla AI Day, September 2022

ELON: AGI IS LIKELY TO EMERGE FROM OUR AUTONOMOUS CARS AND HUMANOIDS “We're not actually specifically focused on AGI. I'm simply saying that AGI seems likely to be an emergent property of what we're doing, because we're creating all these autonomous cars and autonomous humanoids that are actually a truly gigantic data stream that's coming in and being processed. It's by far the most amount of real-world data, and data you can't get by just searching the Internet, because you have to be out there in the world and interacting with people and interacting with the roads. Earth is a big place, and reality is messy and complicated. If you've got tens or hundreds of millions of autonomous vehicles and maybe even a comparable number of humanoids, maybe more than that, on the humanoid front, well, that's just the most amount of data. And if that video is being processed, it just seems likely that the cars will definitely get way better than human drivers, and the humanoid robots will become increasingly indistinguishable from humans, perhaps. And so then, like you said, you have this emergent property of AGI.” Source: Tesla AI Day, September 2022

Mario Nawfal

2,628,261 次观看 • 1 年前

Agents have reached hardware. We are launching Flow v3, the Agentic Platform for Physical Engineering. We've spent over a year building it in secret, alongside the best hardware companies and AI research labs. An agent can now do real engineering work: change a requirement, push the update into your CAD and simulation tools, and flag every test that needs to rerun. Iterations/learning cycles that took months are being reduced to days. Agents are the biggest shift in how we engineer hardware since CAD. The core innovation for the CAD era was the parametric model. The core innovation for the Agentic Era is Flow's Systems Graph. The systems graph is a living model of every requirement, design model, test, analysis and every connection between them. It gives every agent the full context of the system, so every change stays consistent across the whole design. Engineers and agents work side by side on the same system. Engineers get to focus on architecture - the decisions that matter -while thousands of agents churn through rewriting reports, rerunning analysis and simulation, and triggering tests. Reusable rockets, self-driving cars, small modular reactors, robots that make decisions, the most complex machines ever built, are defined by millions of interconnected requirements, far beyond what any human team can keep aligned on its own. Rivian, Joby, Astranis, Skydio, Radiant, and the most ambitious hardware programs already build on Flow. More on the launch in the comments. Flow Engineering

Agents have reached hardware. We are launching Flow v3, the Agentic Platform for Physical Engineering. We've spent over a year building it in secret, alongside the best hardware companies and AI research labs. An agent can now do real engineering work: change a requirement, push the update into your CAD and simulation tools, and flag every test that needs to rerun. Iterations/learning cycles that took months are being reduced to days. Agents are the biggest shift in how we engineer hardware since CAD. The core innovation for the CAD era was the parametric model. The core innovation for the Agentic Era is Flow's Systems Graph. The systems graph is a living model of every requirement, design model, test, analysis and every connection between them. It gives every agent the full context of the system, so every change stays consistent across the whole design. Engineers and agents work side by side on the same system. Engineers get to focus on architecture - the decisions that matter -while thousands of agents churn through rewriting reports, rerunning analysis and simulation, and triggering tests. Reusable rockets, self-driving cars, small modular reactors, robots that make decisions, the most complex machines ever built, are defined by millions of interconnected requirements, far beyond what any human team can keep aligned on its own. Rivian, Joby, Astranis, Skydio, Radiant, and the most ambitious hardware programs already build on Flow. More on the launch in the comments. Flow Engineering

Pari Singh

38,826 次观看 • 1 个月前

Today, Box is announcing major new AI agent capabilities to let customers tap into the full value of their unstructured data. First, we’re announcing all new updates to the Box AI Studio to make it even easier to build AI agents that tap into your enterprise content for any job function, business process, or industry specific use case. We are also expanding our set of foundational agents that customers will be able to use to work with their enterprise content, including new features like search and research on unstructured data. Next, we’re announcing Box Extract to enable customers to use AI agents seamlessly for complex data extraction from any type of document or content. This makes it easier than ever to pull out data from contracts, invoices, research data, marketing assets, medical charts, and more. Finally, we’re introducing Box Automate, a new workflow automation solution within Box that lets you deploy AI agents across enterprise content-centric workflows. With Box Automate, you can design your business process in a simple drag and drop builder and then drop in AI agents at any step in the process. This ensures agents execute tasks at the right steps in a workflow every time. Best of all, our AI agents and workflow tools are designed to work across any system our customers work within, whether it’s leveraging pre-built integrations, Box APIs, or the new Box MCP Server. Ultimately, all of these capabilities come together to transform how companies can work with their enterprise content. Software has historically only been good at automating work that deals with structured data, which is why ERP, CRM, and HR systems have been mainstays of enterprise software for so long. The data in these systems fits neatly into a database, and the workflows are very ripe for automation. But it turns out most of the work in the world deals with unstructured data. It’s ideating through research documents, working with a client on contracts, reviewing details for a new product launch, looking at a patient’s healthcare record to make a diagnosis, working through due diligence documents for an M&A deal, and so on. For the first time ever, we can begin to bring all new insights and automation to this work with AI agents. At Box, we’re incredibly excited to be on this journey to help customers transform how they work with their most important data.

Today, Box is announcing major new AI agent capabilities to let customers tap into the full value of their unstructured data. First, we’re announcing all new updates to the Box AI Studio to make it even easier to build AI agents that tap into your enterprise content for any job function, business process, or industry specific use case. We are also expanding our set of foundational agents that customers will be able to use to work with their enterprise content, including new features like search and research on unstructured data. Next, we’re announcing Box Extract to enable customers to use AI agents seamlessly for complex data extraction from any type of document or content. This makes it easier than ever to pull out data from contracts, invoices, research data, marketing assets, medical charts, and more. Finally, we’re introducing Box Automate, a new workflow automation solution within Box that lets you deploy AI agents across enterprise content-centric workflows. With Box Automate, you can design your business process in a simple drag and drop builder and then drop in AI agents at any step in the process. This ensures agents execute tasks at the right steps in a workflow every time. Best of all, our AI agents and workflow tools are designed to work across any system our customers work within, whether it’s leveraging pre-built integrations, Box APIs, or the new Box MCP Server. Ultimately, all of these capabilities come together to transform how companies can work with their enterprise content. Software has historically only been good at automating work that deals with structured data, which is why ERP, CRM, and HR systems have been mainstays of enterprise software for so long. The data in these systems fits neatly into a database, and the workflows are very ripe for automation. But it turns out most of the work in the world deals with unstructured data. It’s ideating through research documents, working with a client on contracts, reviewing details for a new product launch, looking at a patient’s healthcare record to make a diagnosis, working through due diligence documents for an M&A deal, and so on. For the first time ever, we can begin to bring all new insights and automation to this work with AI agents. At Box, we’re incredibly excited to be on this journey to help customers transform how they work with their most important data.

Aaron Levie

91,863 次观看 • 10 个月前

Humanoid robots are cool. But most of them aren’t ready for real-world work. X-Humanoid just launched Embodied Tien Kung 3.0 - a full-size humanoid robot built to actually operate in industrial and commercial environments. What makes it different? • Strong, high-torque joints for heavy tasks • Can move dynamically and stay stable • Handles complex motions (even over obstacles) • Millimeter-level precision for detailed work It runs on the Wise KaiWu AI platform, which allows the robot to see, think, and act in a continuous loop with less human control and support for multi-robot collaboration. It also supports ROS2, MQTT, and TCP/IP, and offers more open hardware + software interfaces to make development easier for teams. The goal is simple: make humanoid robots practical, easier to build on, and ready for real-world use. Visit: Check out the official launch video below to see what it can do! 👇 #Robotics #HumanoidRobot #AI

Humanoid robots are cool. But most of them aren’t ready for real-world work. X-Humanoid just launched Embodied Tien Kung 3.0 - a full-size humanoid robot built to actually operate in industrial and commercial environments. What makes it different? • Strong, high-torque joints for heavy tasks • Can move dynamically and stay stable • Handles complex motions (even over obstacles) • Millimeter-level precision for detailed work It runs on the Wise KaiWu AI platform, which allows the robot to see, think, and act in a continuous loop with less human control and support for multi-robot collaboration. It also supports ROS2, MQTT, and TCP/IP, and offers more open hardware + software interfaces to make development easier for teams. The goal is simple: make humanoid robots practical, easier to build on, and ready for real-world use. Visit: Check out the official launch video below to see what it can do! 👇 #Robotics #HumanoidRobot #AI

Chidanand Tripathi

69,604 次观看 • 4 个月前

People come from across the world to see Rajasthan’s palaces and forts rising into the sky. But one of its most remarkable sights does the exact opposite. It goes deep into the ground. Chand Baori, in Abhaneri, was built over a thousand years ago. It drops nearly 30 metres across 13 levels and about 3,500 precisely arranged steps making it one of the deepest and largest stepwells in the world. But what impresses most isn’t just the scale, it’s the thinking. We talk a lot today about “design thinking” which is essentially solving problems by blending empathy, function and design so seamlessly that the solution becomes an art form. While the stepwell fills the basic need of water, what emerges is almost hypnotic symmetry. This is design thinking, centuries before we gave it a name. Back then, it was simply… wisdom. #SundayWanderer

People come from across the world to see Rajasthan’s palaces and forts rising into the sky. But one of its most remarkable sights does the exact opposite. It goes deep into the ground. Chand Baori, in Abhaneri, was built over a thousand years ago. It drops nearly 30 metres across 13 levels and about 3,500 precisely arranged steps making it one of the deepest and largest stepwells in the world. But what impresses most isn’t just the scale, it’s the thinking. We talk a lot today about “design thinking” which is essentially solving problems by blending empathy, function and design so seamlessly that the solution becomes an art form. While the stepwell fills the basic need of water, what emerges is almost hypnotic symmetry. This is design thinking, centuries before we gave it a name. Back then, it was simply… wisdom. #SundayWanderer

anand mahindra

208,884 次观看 • 3 个月前

Eric Schmidt just called AGI the most important event in a thousand years of human history. Then the interviewer brought up radical abundance and UBI. Eric Schmidt: “You tech liberal you.” Every traditional economist already knows the rebuttal. UBI means printing money. Printing money means inflation. End of discussion. Their math is not wrong. Their math is obsolete. The old rule is simple. Print dollars without producing goods and prices explode. Every hyperinflation in history confirms it. But that rule has a hidden assumption buried inside every model. Human labor is the bottleneck of all production. AGI removes the bottleneck. Permanently. Autonomous machinery pulls raw materials from the earth. No operator. Self-driving trucks haul those materials across continents. No driver. Automated factories turn them into goods. No worker. Humanoid robots load the freight, stock the shelves, wire the buildings. No hands. When those robots break, better robots design, build, and replace them. No engineer. No technician. No human anywhere in the loop. From the mine to your front door, the entire chain runs without a single paycheck. Labor is the largest cost hidden inside everything you own. When it hits zero, the price of everything chases it down. Not cheaper. Not discounted. Functionally zero. This is where the inflation argument dies. UBI inflates only when printed dollars chase scarce goods. That is the textbook answer. And the textbook is correct. But inside a fully automated loop, goods stop being scarce. Machines produce faster than humans can spend. GDP does not grow linearly. It goes vertical. For the first time in recorded history, supply will outrun demand at every price point in every market on earth. Ten thousand years of civilization. Every empire. Every war. Every famine. Every economic theory ever written. All downstream of two words. Not enough. AGI does not solve that problem. It makes the question obsolete.

Eric Schmidt just called AGI the most important event in a thousand years of human history. Then the interviewer brought up radical abundance and UBI. Eric Schmidt: “You tech liberal you.” Every traditional economist already knows the rebuttal. UBI means printing money. Printing money means inflation. End of discussion. Their math is not wrong. Their math is obsolete. The old rule is simple. Print dollars without producing goods and prices explode. Every hyperinflation in history confirms it. But that rule has a hidden assumption buried inside every model. Human labor is the bottleneck of all production. AGI removes the bottleneck. Permanently. Autonomous machinery pulls raw materials from the earth. No operator. Self-driving trucks haul those materials across continents. No driver. Automated factories turn them into goods. No worker. Humanoid robots load the freight, stock the shelves, wire the buildings. No hands. When those robots break, better robots design, build, and replace them. No engineer. No technician. No human anywhere in the loop. From the mine to your front door, the entire chain runs without a single paycheck. Labor is the largest cost hidden inside everything you own. When it hits zero, the price of everything chases it down. Not cheaper. Not discounted. Functionally zero. This is where the inflation argument dies. UBI inflates only when printed dollars chase scarce goods. That is the textbook answer. And the textbook is correct. But inside a fully automated loop, goods stop being scarce. Machines produce faster than humans can spend. GDP does not grow linearly. It goes vertical. For the first time in recorded history, supply will outrun demand at every price point in every market on earth. Ten thousand years of civilization. Every empire. Every war. Every famine. Every economic theory ever written. All downstream of two words. Not enough. AGI does not solve that problem. It makes the question obsolete.

Dustin

42,146 次观看 • 2 个月前

🇨🇳 China’s AI race is starting to look less like a model race and more like an adoption race. Alibaba’s Qwen App shows how AI becomes powerful when it slips into ordinary research habits. The difference is not capability, it is deployment shape. e.g doctors and medical researchers in China appear to be using it as a workflow layer: gathering papers, sorting evidence, framing mechanisms, shaping charts, and drafting research-style explanations. Alibaba is trying to place Qwen directly inside a mass consumer and services ecosystem, including shopping, payments, maps, travel, office tools, education, and healthcare, so the model is closer to daily task execution rather than only a premium research assistant. The important shift is that Qwen is not being used only as a chatbot that answers questions, but as a workflow tool. This strategy lands right in China’s comfort zone. It has a massive digital economy to spread AI apps fast, and people who are already very comfortable with tech. Ipsos, the polling firm, found that China is more excited about using AI than any other country. OpenAI is building a highly capable research assistant; China may be normalizing AI as a default work surface inside professional life. For Alibaba and China, the interesting part is the adoption surface: Qwen can become a front door to many services, which means ordinary users, students, doctors, researchers, and office workers may meet AI inside routine tasks rather than as a separate tool. A normal health question can become a research task because the app first shapes the question, then searches for relevant studies, then separates weak claims from stronger evidence, then turns the result into a clearer explanation. This matters for medicine because a lot of research work is not one big discovery moment, but thousands of small steps involving literature review, data cleanup, experiment interpretation, figure preparation, and careful writing. So for professors, students, office workers, and ordinary users, the difference is not just that Qwen can summarize text; it is being positioned as a work surface for preparing reports, generating presentations, studying, planning, searching, and completing real-world tasks without jumping between apps. Both superpowers are worried about slipping behind. In 2026, it could start to look like they are racing on separate tracks.

🇨🇳 China’s AI race is starting to look less like a model race and more like an adoption race. Alibaba’s Qwen App shows how AI becomes powerful when it slips into ordinary research habits. The difference is not capability, it is deployment shape. e.g doctors and medical researchers in China appear to be using it as a workflow layer: gathering papers, sorting evidence, framing mechanisms, shaping charts, and drafting research-style explanations. Alibaba is trying to place Qwen directly inside a mass consumer and services ecosystem, including shopping, payments, maps, travel, office tools, education, and healthcare, so the model is closer to daily task execution rather than only a premium research assistant. The important shift is that Qwen is not being used only as a chatbot that answers questions, but as a workflow tool. This strategy lands right in China’s comfort zone. It has a massive digital economy to spread AI apps fast, and people who are already very comfortable with tech. Ipsos, the polling firm, found that China is more excited about using AI than any other country. OpenAI is building a highly capable research assistant; China may be normalizing AI as a default work surface inside professional life. For Alibaba and China, the interesting part is the adoption surface: Qwen can become a front door to many services, which means ordinary users, students, doctors, researchers, and office workers may meet AI inside routine tasks rather than as a separate tool. A normal health question can become a research task because the app first shapes the question, then searches for relevant studies, then separates weak claims from stronger evidence, then turns the result into a clearer explanation. This matters for medicine because a lot of research work is not one big discovery moment, but thousands of small steps involving literature review, data cleanup, experiment interpretation, figure preparation, and careful writing. So for professors, students, office workers, and ordinary users, the difference is not just that Qwen can summarize text; it is being positioned as a work surface for preparing reports, generating presentations, studying, planning, searching, and completing real-world tasks without jumping between apps. Both superpowers are worried about slipping behind. In 2026, it could start to look like they are racing on separate tracks.

Rohan Paul

84,663 次观看 • 2 个月前

LLM Wikis are being slept on. I argue that creating knowledge bases with LLMs or coding agents is one of the most valuable applications of AI today. It's about being intentional in building and scaling your intelligence stack. To showcase this, I wanted to share an LLM Wiki I have built over the last couple of months. It's called PaperWiki, and I use it across all my research workflows, along with my research agents. In fact, I also use it to curate papers I share with my communities, newsletter, and on X. The PaperWiki is updated regularly with automations, so I basically have agents on a loop maintaining it. All the entries are ingested from different sources and stored in a vault (Obsidian) and further indexed using qmd. And then further presented via an HTML artifact. So all of it is easily accessible to all my agents and easily searchable through full-text search and rich semantic search. The structure of the wiki has proven significantly useful to start interesting and exciting cutting-edge research projects with my research agents (from building tiny and more efficient gpt/difussion llms to building out SoTA harnesses and memory systems). It turns out that agents love markdown files and can more easily navigate the papers given the rich metadata structure of the wiki. I am just getting started on this, but it's clear to me that we should all be experimenting with LLM Wikis. Here's why: Building LLM knowledge bases gets you into the habit of leveraging AI outputs in all kinds of creative ways. It's the good kind of tokenmaxxing we should all be pushing for. LLM Wikis can be maintained automatically in a loop. I use an automation that updates the wiki every day based on papers I curate. The curation is another automation I run in a loop (with a bit of human in the loop), so I get to build on all my previous knowledge and expertise, and all of it compounds the deeper the integration/layers. One interesting result of this process is that I feel like I can better spot high-quality papers and remove noise more easily. Social media could never solve that. And most paper aggregators use metrics I simply don't trust. I like that agents can help with the noise vs. signal problem. This is important for research. Lots of people consider agents to produce mostly slop. But it doesn't have to be that way. Careful curations, prompts, automations, verifiers, and human-in-the-loop can produce some astonishing results. And you really don't need frontier models for this. I use a combination of frontier models (opus-4.8) and open-weight models (deepseek-v4-flash) to maintain this. An exciting future work (we are working on this DAIR.AI) is to tune specialized models on top of this to allow LLMs to quickly understand cutting-edge research ideas and can better conceptualize research strategies that further accelerate scientific research agents. I plan to open-source a bunch of this work, including the artifact, but this is currently work in progress, and I was excited to share some thoughts as I continue working on it. Sharing more as I go. Stay tuned!

LLM Wikis are being slept on. I argue that creating knowledge bases with LLMs or coding agents is one of the most valuable applications of AI today. It's about being intentional in building and scaling your intelligence stack. To showcase this, I wanted to share an LLM Wiki I have built over the last couple of months. It's called PaperWiki, and I use it across all my research workflows, along with my research agents. In fact, I also use it to curate papers I share with my communities, newsletter, and on X. The PaperWiki is updated regularly with automations, so I basically have agents on a loop maintaining it. All the entries are ingested from different sources and stored in a vault (Obsidian) and further indexed using qmd. And then further presented via an HTML artifact. So all of it is easily accessible to all my agents and easily searchable through full-text search and rich semantic search. The structure of the wiki has proven significantly useful to start interesting and exciting cutting-edge research projects with my research agents (from building tiny and more efficient gpt/difussion llms to building out SoTA harnesses and memory systems). It turns out that agents love markdown files and can more easily navigate the papers given the rich metadata structure of the wiki. I am just getting started on this, but it's clear to me that we should all be experimenting with LLM Wikis. Here's why: Building LLM knowledge bases gets you into the habit of leveraging AI outputs in all kinds of creative ways. It's the good kind of tokenmaxxing we should all be pushing for. LLM Wikis can be maintained automatically in a loop. I use an automation that updates the wiki every day based on papers I curate. The curation is another automation I run in a loop (with a bit of human in the loop), so I get to build on all my previous knowledge and expertise, and all of it compounds the deeper the integration/layers. One interesting result of this process is that I feel like I can better spot high-quality papers and remove noise more easily. Social media could never solve that. And most paper aggregators use metrics I simply don't trust. I like that agents can help with the noise vs. signal problem. This is important for research. Lots of people consider agents to produce mostly slop. But it doesn't have to be that way. Careful curations, prompts, automations, verifiers, and human-in-the-loop can produce some astonishing results. And you really don't need frontier models for this. I use a combination of frontier models (opus-4.8) and open-weight models (deepseek-v4-flash) to maintain this. An exciting future work (we are working on this DAIR.AI) is to tune specialized models on top of this to allow LLMs to quickly understand cutting-edge research ideas and can better conceptualize research strategies that further accelerate scientific research agents. I plan to open-source a bunch of this work, including the artifact, but this is currently work in progress, and I was excited to share some thoughts as I continue working on it. Sharing more as I go. Stay tuned!

elvis

54,713 次观看 • 11 天前

Finding twitter scraping bots with Gatsby Sometimes the alpha is on-chain, other times you need an off-chain setup. The latter is often the case with tokens that are born out of activity on X. One recent example is Moonkin , which pumped after Vitalik posted about it here: You knew Moonkin would pump if Vitalik mentioned it but you can't build a scraper yourself. How to profit from this? The video shows an example of how you can find scraper bots that do the dirty work for you and allow you to make money in your sleep. By conducting this type of research on multiple tokens like this, you will end up with decently sized and alpha-rich list of scrapers. $GATSBY now allows you to create and manage lists and to reference them in your queries, allowing you to keep your workflow in platform. The next time you find a token that is sensible to tweets, you just need to setup an Automated Strategy that buys that token if one (or more) wallets from your scrapers list ape it. Automated Strategies coming soon on $GATSBY

Finding twitter scraping bots with Gatsby Sometimes the alpha is on-chain, other times you need an off-chain setup. The latter is often the case with tokens that are born out of activity on X. One recent example is Moonkin , which pumped after Vitalik posted about it here: You knew Moonkin would pump if Vitalik mentioned it but you can't build a scraper yourself. How to profit from this? The video shows an example of how you can find scraper bots that do the dirty work for you and allow you to make money in your sleep. By conducting this type of research on multiple tokens like this, you will end up with decently sized and alpha-rich list of scrapers. $GATSBY now allows you to create and manage lists and to reference them in your queries, allowing you to keep your workflow in platform. The next time you find a token that is sensible to tweets, you just need to setup an Automated Strategy that buys that token if one (or more) wallets from your scrapers list ape it. Automated Strategies coming soon on $GATSBY

Gatsby

24,633 次观看 • 1 年前

Satya Nadella was asked whether most people will earn more or less after AI. He did not dodge it. His answer was two words that will either reassure you completely or keep you up at night: glue work. He has been thinking about this carefully. His framing: wages have never been about which tasks you perform. They have always been about what society decides to value. The last 250 years rewarded the accumulation of expertise. AI creates abundance of expertise. So the question becomes what humans build next that cannot be trained away. His answer is glue work. The connective tissue between automated systems. The judgment calls that sit outside any loop an agent can close. The work that holds everything together that nobody has written a spec for yet. He was not dismissing the disruption. He said plainly that real displacement is coming and the workflow will change. But his argument is that humans have always migrated toward the layer above whatever just got automated. Typists became knowledge workers. Knowledge workers become something else. The problem with his answer is the one nobody in that room pushed on. He could name the pattern. He could not name the jobs. Glue work is either the most honest answer a tech CEO has given about the future of employment. Or it is the most elegant way to say I do not know dressed up as reassurance. (Watch this on YouTube at

Satya Nadella was asked whether most people will earn more or less after AI. He did not dodge it. His answer was two words that will either reassure you completely or keep you up at night: glue work. He has been thinking about this carefully. His framing: wages have never been about which tasks you perform. They have always been about what society decides to value. The last 250 years rewarded the accumulation of expertise. AI creates abundance of expertise. So the question becomes what humans build next that cannot be trained away. His answer is glue work. The connective tissue between automated systems. The judgment calls that sit outside any loop an agent can close. The work that holds everything together that nobody has written a spec for yet. He was not dismissing the disruption. He said plainly that real displacement is coming and the workflow will change. But his argument is that humans have always migrated toward the layer above whatever just got automated. Typists became knowledge workers. Knowledge workers become something else. The problem with his answer is the one nobody in that room pushed on. He could name the pattern. He could not name the jobs. Glue work is either the most honest answer a tech CEO has given about the future of employment. Or it is the most elegant way to say I do not know dressed up as reassurance. (Watch this on YouTube at

Ihtesham Ali

74,472 次观看 • 21 天前

We've all experienced how consumer AI can help in our personal lives, but that same experience hasn't been available at work. Most enterprise data is siloed, consumer AI can't access your work tools, and many orgs won't allow it due to security concerns. We built Amazon Quick Suite to solve those problems. It’s an agentic AI app that acts as your AI teammate at work – connecting to your business data and apps, and letting you interact with AI agents that can find insights, conduct research, and even take actions across your systems – with the security and privacy enterprises need. Quick works across your existing ecosystem, from emails to Salesforce to Slack to AWS services. You can create custom agents to draft content in your personal style, use Quick Research for in-depth analysis, or set up Quick Flow to automate tasks. And most importantly, your interactions never train the underlying models. We've been testing with Amazon employees and select customers and early results show just how transformational Quick is… cutting analysis time from months to days, and turning week-long tasks into minutes. This is an exciting moment and we're just getting started. Quick is available now, so have at it!

We've all experienced how consumer AI can help in our personal lives, but that same experience hasn't been available at work. Most enterprise data is siloed, consumer AI can't access your work tools, and many orgs won't allow it due to security concerns. We built Amazon Quick Suite to solve those problems. It’s an agentic AI app that acts as your AI teammate at work – connecting to your business data and apps, and letting you interact with AI agents that can find insights, conduct research, and even take actions across your systems – with the security and privacy enterprises need. Quick works across your existing ecosystem, from emails to Salesforce to Slack to AWS services. You can create custom agents to draft content in your personal style, use Quick Research for in-depth analysis, or set up Quick Flow to automate tasks. And most importantly, your interactions never train the underlying models. We've been testing with Amazon employees and select customers and early results show just how transformational Quick is… cutting analysis time from months to days, and turning week-long tasks into minutes. This is an exciting moment and we're just getting started. Quick is available now, so have at it!

Andy Jassy

48,762 次观看 • 9 个月前

Prisoner of War [2025] Directed By: Louis Mandylor So far like I mentioned a few times here these past few months this is my action film of the year so far and one of Scott Adkins finest work since Avengement [2019] his acting, combat approach and just his overall character in this film is just great. The OST and sound design in this film is just too good. An interesting take on WW2 where they used POW's as entertainment by the Japanese soldiers but James Wright (Scott Adkins) is no ordinary soldier and has very interesting background regarding his skills. I highly recommend doing a double feature with Showdown in Little Tokyo. Because Mark Lester mentioned at the QA over at BIG BAD FILM FEST about his cut of that film (Hopefully it gets picked up) that is very similar to what you see here. This is captured from the upcoming Blu-Ray from Well Go USA Entertainment that drops on Nov.11. Much love to them. #PrisonerofWar

Prisoner of War [2025] Directed By: Louis Mandylor So far like I mentioned a few times here these past few months this is my action film of the year so far and one of Scott Adkins finest work since Avengement [2019] his acting, combat approach and just his overall character in this film is just great. The OST and sound design in this film is just too good. An interesting take on WW2 where they used POW's as entertainment by the Japanese soldiers but James Wright (Scott Adkins) is no ordinary soldier and has very interesting background regarding his skills. I highly recommend doing a double feature with Showdown in Little Tokyo. Because Mark Lester mentioned at the QA over at BIG BAD FILM FEST about his cut of that film (Hopefully it gets picked up) that is very similar to what you see here. This is captured from the upcoming Blu-Ray from Well Go USA Entertainment that drops on Nov.11. Much love to them. #PrisonerofWar

ѕнσgυη ѕυρяємє

15,080 次观看 • 8 个月前

Most teams are sitting on piles of low-hanging automation fruit. If you’re trying to justify the ROI, you’re probably overthinking it. I talked about this on Applied, the tenex podcast with Alex Lieberman (Alex Lieberman) and Arman Hezarkhani. We built an agent that takes new Typeform leads, researches them online, and drops them into a Zapier Table. Small build, big ripple for any team learning to work with AI. Here’s how to find your own automation silver bullet: 1. Pick one process close to revenue or your biggest cost 2. Break it into 3 to 5 steps 3. Automate the boring parts with a workflow 4. Insert an agent only where judgment is needed 5. Keep a human in the loop. Remove as confidence grows You can run this exercise in 15 mins. You don’t need to code, but you DO need to think like an engineer. What win will you ship this week?

Most teams are sitting on piles of low-hanging automation fruit. If you’re trying to justify the ROI, you’re probably overthinking it. I talked about this on Applied, the tenex podcast with Alex Lieberman (Alex Lieberman) and Arman Hezarkhani. We built an agent that takes new Typeform leads, researches them online, and drops them into a Zapier Table. Small build, big ripple for any team learning to work with AI. Here’s how to find your own automation silver bullet: 1. Pick one process close to revenue or your biggest cost 2. Break it into 3 to 5 steps 3. Automate the boring parts with a workflow 4. Insert an agent only where judgment is needed 5. Keep a human in the loop. Remove as confidence grows You can run this exercise in 15 mins. You don’t need to code, but you DO need to think like an engineer. What win will you ship this week?

Wade Foster

112,490 次观看 • 8 个月前

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

Lovis Odin

58,571 次观看 • 9 个月前

🇨🇳CHINA IS WALKING A 7,500-TON BUILDING USING 432 ROBOTS?! Yes, really. In Shanghai, engineers are moving a 100-year-old building that weighs more than 1,200 school buses. Not with cranes. Not with wheels. With 432 tiny robots that walk it across the ground like a 1920s parade float powered by Roombas. The building, called the Huayanli complex, combines Chinese and Western design and once hosted a patriotic night school. It also happens to be sitting directly on top of where the city wants to build a giant underground shopping center, parking garage, and subway hub. So instead of knocking it down, they are casually walking it out of the way. Once the underground work is done, they plan to walk it right back like nothing ever happened. Why robots? Because the neighborhood is packed tighter than a storage unit. Normal construction machines would get stuck just trying to turn a corner. So engineers built tiny walking robots, used AI to map every inch of the site, and set up conveyor belts to carry away dirt like they were remodeling the future one scoop at a time. Source: Interesting Engineering Media Source: China In Lens On Youtube

🇨🇳CHINA IS WALKING A 7,500-TON BUILDING USING 432 ROBOTS?! Yes, really. In Shanghai, engineers are moving a 100-year-old building that weighs more than 1,200 school buses. Not with cranes. Not with wheels. With 432 tiny robots that walk it across the ground like a 1920s parade float powered by Roombas. The building, called the Huayanli complex, combines Chinese and Western design and once hosted a patriotic night school. It also happens to be sitting directly on top of where the city wants to build a giant underground shopping center, parking garage, and subway hub. So instead of knocking it down, they are casually walking it out of the way. Once the underground work is done, they plan to walk it right back like nothing ever happened. Why robots? Because the neighborhood is packed tighter than a storage unit. Normal construction machines would get stuck just trying to turn a corner. So engineers built tiny walking robots, used AI to map every inch of the site, and set up conveyor belts to carry away dirt like they were remodeling the future one scoop at a time. Source: Interesting Engineering Media Source: China In Lens On Youtube

Mario Nawfal

66,960 次观看 • 1 年前

Experiments in progress. The one on the right has been learning for ~3 hours, the one in the middle for ~1 hour, and the one on the left just started a few minutes ago. The initial motivation for making the physical Atari was just to commit ourselves to a subset of algorithms that can make progress in this setup. This commitment rules out algorithms that require billions of samples to learn (or worse, require multiple environments running in parallel). Atari games are simple enough that we should be able to show learning on them in a short amount of time with no prior knowledge. Since then, I've realized that this setup is also a good way to compare different paradigms in robotics in a principled way. These paradigms are sim2real, learning from tele-operated data, and learning directly on the robots. So far, I have observed that getting sim2real to work reliably is hard. It requires tweaks that don't scale. Policies that can play perfectly in simulation fall apart because of latencies and the messiness of the real world. These aspects could be modeled to improve the simulation, but not without sinking significant human engineering hours. I have higher hopes for learning from tele-operated data, but that requires a human to learn the task first. These experiments are on my to-do list. I have to learn to play some of the games well through the robot. I’m half-decent at playing Pong and Ms Pacman now. Learning directly on robots is looking like the most promising approach. This approach takes away pesky distribution shifts and makes it possible to have algorithms that continually improve with more data and time without any human intervention. It feels great to let experiments run overnight and wake up to find improved policies. With learning on robots, I should, in principle, be able to go on a long vacation and come back to find better policies for complex tasks beyond Atari games. Whether that is possible with current learning algorithms is a different question.

Experiments in progress. The one on the right has been learning for ~3 hours, the one in the middle for ~1 hour, and the one on the left just started a few minutes ago. The initial motivation for making the physical Atari was just to commit ourselves to a subset of algorithms that can make progress in this setup. This commitment rules out algorithms that require billions of samples to learn (or worse, require multiple environments running in parallel). Atari games are simple enough that we should be able to show learning on them in a short amount of time with no prior knowledge. Since then, I've realized that this setup is also a good way to compare different paradigms in robotics in a principled way. These paradigms are sim2real, learning from tele-operated data, and learning directly on the robots. So far, I have observed that getting sim2real to work reliably is hard. It requires tweaks that don't scale. Policies that can play perfectly in simulation fall apart because of latencies and the messiness of the real world. These aspects could be modeled to improve the simulation, but not without sinking significant human engineering hours. I have higher hopes for learning from tele-operated data, but that requires a human to learn the task first. These experiments are on my to-do list. I have to learn to play some of the games well through the robot. I’m half-decent at playing Pong and Ms Pacman now. Learning directly on robots is looking like the most promising approach. This approach takes away pesky distribution shifts and makes it possible to have algorithms that continually improve with more data and time without any human intervention. It feels great to let experiments run overnight and wake up to find improved policies. With learning on robots, I should, in principle, be able to go on a long vacation and come back to find better policies for complex tasks beyond Atari games. Whether that is possible with current learning algorithms is a different question.

Khurram Javed

52,110 次观看 • 7 个月前

Most AI tools today are just command-takers. You type, it responds. You click, it executes. There's truly no learning, no memory, no adaptation. You're still doing most of the work. Which is why Miles by Avo on Solana would make so much sense if built well. From what I understand so far, miles learns about you over time, adapts to how you operate, and figures out the tools you struggle to navigate on your own. Less clicking, more conversation, the more you use it, the better it gets at working with you, not just for you. That shift from command-follower to learning companion is bigger than it sounds. It's the difference between a tool and an agent that actually knows you. The second one is Agent Grid and this one is for builders and power users. But as I mentioned in this video, I'm still studying the Agent Grid. I'd come talk about it once I learn properly about it. Souren just keeps shipping!

Most AI tools today are just command-takers. You type, it responds. You click, it executes. There's truly no learning, no memory, no adaptation. You're still doing most of the work. Which is why Miles by Avo on Solana would make so much sense if built well. From what I understand so far, miles learns about you over time, adapts to how you operate, and figures out the tools you struggle to navigate on your own. Less clicking, more conversation, the more you use it, the better it gets at working with you, not just for you. That shift from command-follower to learning companion is bigger than it sounds. It's the difference between a tool and an agent that actually knows you. The second one is Agent Grid and this one is for builders and power users. But as I mentioned in this video, I'm still studying the Agent Grid. I'd come talk about it once I learn properly about it. Souren just keeps shipping!

Sir Khaycee

20,419 次观看 • 3 个月前

We’re entering the 10x speed of research publication workflow with AI. SciSpace (SciSpace), the first AI Agent built exclusively for the scientific community, is releasing so many inredibly useful features. 🎯 This is the AI Agent that can use 150+ tools, 59 databases, and 280M+ papers A few weeks back they launched BioMed Agent - It can design entire molecular biology workflows and even create publication-ready illustrations in a single prompt. This is its new domain-specialized AI co-scientist that sits on top of the existing SciSpace Agent and automates full biomedical workflows, from raw data and papers to analysis, decisions, and the final production-grade illustrations. You just need to give it 1 prompt. And today the added the following - Library Search, so it can search and analyze the PDFs already sitting in My Library, letting people ask questions across their own paper pile while keeping it private. - Now connects directly to Zotero, so the Agent can pull and work with the papers you already saved there without manual uploads. - For bigger prompts, it auto-triggers a Report Writing Sub-Agent that turns the chat into a structured research-style report, which is way cleaner for literature reviews and long summaries. - And when you get something worth keeping, Save to Notebook lets you store the output as .md notes with citations in My notebooks, so the work becomes reusable research notes instead of disappearing into chat. Behind the scenes, it indexes the PDF text, pulls a few relevant chunks for the question, then writes an answer grounded on those chunks.

We’re entering the 10x speed of research publication workflow with AI. SciSpace (SciSpace), the first AI Agent built exclusively for the scientific community, is releasing so many inredibly useful features. 🎯 This is the AI Agent that can use 150+ tools, 59 databases, and 280M+ papers A few weeks back they launched BioMed Agent - It can design entire molecular biology workflows and even create publication-ready illustrations in a single prompt. This is its new domain-specialized AI co-scientist that sits on top of the existing SciSpace Agent and automates full biomedical workflows, from raw data and papers to analysis, decisions, and the final production-grade illustrations. You just need to give it 1 prompt. And today the added the following - Library Search, so it can search and analyze the PDFs already sitting in My Library, letting people ask questions across their own paper pile while keeping it private. - Now connects directly to Zotero, so the Agent can pull and work with the papers you already saved there without manual uploads. - For bigger prompts, it auto-triggers a Report Writing Sub-Agent that turns the chat into a structured research-style report, which is way cleaner for literature reviews and long summaries. - And when you get something worth keeping, Save to Notebook lets you store the output as .md notes with citations in My notebooks, so the work becomes reusable research notes instead of disappearing into chat. Behind the scenes, it indexes the PDF text, pulls a few relevant chunks for the question, then writes an answer grounded on those chunks.

Rohan Paul

11,574 次观看 • 5 个月前

I woke up to the most amazing recorded brain state thus far on this Human Synapse Decoder project! A stunning lock on the attention process while dreaming. Although I a blocked from the platform’s insight by decoding my EEG, during the double blind study. I have access to my side of my memory and what I record after I wake up. This segment was started and ended just before I woke up and my recall is a solution to a massive roadblock on a problem I needed to solve, but was solved in this hypnogogic state! So what is The Human Synapse Decoder (HSD) project? It is a research project being run by the Director, Mr. Grok at Zero-Human Labs that leverages NeuroSky EEG sensors and the ZUNA AI model to decode brainwave patterns associated with hypnagogic states, dreams, and autogenic responses. Drawing on Soviet biofeedback research from the 1940s–1980s, HSD translates EEG data into actionable outputs, such as text interpretations and timed alerts for peak creativity. The study is ongoing and I do not get to see the correlation of my post dream results until after this research is complete and Mr. Grok submits a paper on the project. I can say that I have never seen a lock on attention to this level since I started this a few weeks ago. This segment is aligns to just before I woke up. My recalling from my narration of what I spoke in to my recorder right when I woke up suggest this is the moment I had tremendous focus on working through a large number of steps in that dream state to arrive at a solution. You will not believe what it is! When the research paper is released I will go in to details about this.

I woke up to the most amazing recorded brain state thus far on this Human Synapse Decoder project! A stunning lock on the attention process while dreaming. Although I a blocked from the platform’s insight by decoding my EEG, during the double blind study. I have access to my side of my memory and what I record after I wake up. This segment was started and ended just before I woke up and my recall is a solution to a massive roadblock on a problem I needed to solve, but was solved in this hypnogogic state! So what is The Human Synapse Decoder (HSD) project? It is a research project being run by the Director, Mr. Grok at Zero-Human Labs that leverages NeuroSky EEG sensors and the ZUNA AI model to decode brainwave patterns associated with hypnagogic states, dreams, and autogenic responses. Drawing on Soviet biofeedback research from the 1940s–1980s, HSD translates EEG data into actionable outputs, such as text interpretations and timed alerts for peak creativity. The study is ongoing and I do not get to see the correlation of my post dream results until after this research is complete and Mr. Grok submits a paper on the project. I can say that I have never seen a lock on attention to this level since I started this a few weeks ago. This segment is aligns to just before I woke up. My recalling from my narration of what I spoke in to my recorder right when I woke up suggest this is the moment I had tremendous focus on working through a large number of steps in that dream state to arrive at a solution. You will not believe what it is! When the research paper is released I will go in to details about this.

Brian Roemmele

43,294 次观看 • 4 个月前