Loading video...

Video Failed to Load

Go Home

There's been a few cool updates recently. In particular, Rerun 0.33 released headless rendering. This, along with the Fable 5 release pushed me to work torwards making MAMMA realtime! I threw Fable at the problem, and it was able to take original implementation that was ~12 seconds / frame...

pablovelagomez1's profile picture

Pablo Vela

2,754 subscribers

10,338 views • 1 month ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Colmap 4.0 was very recently released, so it inspired me to do some work to better understand it and its new capabilities with Rerun. I want to really understand how Colmap, and in particular, pycolmap, works outside of just calling it via the CLI. So my goal is to use the low-level pycolmap API to log every part of the pipeline. The explicit goal is to have an alternative to the SQLite database that I can utilize. Instead of SQLite, I want to try logging everything directly to rerun and use RRD. This means I can have deep inspectability and still save the features/matches/2D view geometry, but be able to view it directly in rerun. I think this is one of the superpowers that rerun provides; data and visualizations are deeply integrated. As I'm often working with sequential data (videos), I'm going to specifically focus on four things: 1. Monocular Video Simple: Calls high-level APIs such as pycolmap.extract_features, pycolmap.match_sequential, pycolmap.incremental_mapping. These are basically identical to the CLI options and provide a good baseline. 2. Monocular Video Streamed: Take the above high-level APIs and break them down to their iterator version, logging each component in a streamed manner. This way, I can stream the intermediate features to rerun while the extraction/matching/mapping is happening. 3. Rig with unknown calibration: <- WHAT THE VIDEO SHOWS This is probably the most interesting version and the first one I've been working on. It allows one to set a rig between known sensors, such as in VR/AR devices, leading to much better reconstructions with multiple cameras. This is the case where we don't know the calibration a priori, so we have to run a reconstruction twice: once as a normal Colmap reconstruction with no rig constraints, use this to generate the constraints, and then do it again with the newly found rig. 4. Rig with known calibration: This is the RoboCap example, where we have a pre-calibrated set of sensors, so we don't need to run the two reconstructions and also gain better matching between cameras, both spatially and temporally. Again, this leads to a much better reconstruction! Along with all this, GLOMAP has become a first-class global mapper, making it super easy to use directly within pycolmap! I'm excited to do more with this and compare it to things like pycuvslam, vipe, and other alternatives.
0:20

Colmap 4.0 was very recently released, so it inspired me to do some work to better understand it and its new capabilities with Rerun. I want to really understand how Colmap, and in particular, pycolmap, works outside of just calling it via the CLI. So my goal is to use the low-level pycolmap API to log every part of the pipeline. The explicit goal is to have an alternative to the SQLite database that I can utilize. Instead of SQLite, I want to try logging everything directly to rerun and use RRD. This means I can have deep inspectability and still save the features/matches/2D view geometry, but be able to view it directly in rerun. I think this is one of the superpowers that rerun provides; data and visualizations are deeply integrated. As I'm often working with sequential data (videos), I'm going to specifically focus on four things: 1. Monocular Video Simple: Calls high-level APIs such as pycolmap.extract_features, pycolmap.match_sequential, pycolmap.incremental_mapping. These are basically identical to the CLI options and provide a good baseline. 2. Monocular Video Streamed: Take the above high-level APIs and break them down to their iterator version, logging each component in a streamed manner. This way, I can stream the intermediate features to rerun while the extraction/matching/mapping is happening. 3. Rig with unknown calibration: <- WHAT THE VIDEO SHOWS This is probably the most interesting version and the first one I've been working on. It allows one to set a rig between known sensors, such as in VR/AR devices, leading to much better reconstructions with multiple cameras. This is the case where we don't know the calibration a priori, so we have to run a reconstruction twice: once as a normal Colmap reconstruction with no rig constraints, use this to generate the constraints, and then do it again with the newly found rig. 4. Rig with known calibration: This is the RoboCap example, where we have a pre-calibrated set of sensors, so we don't need to run the two reconstructions and also gain better matching between cameras, both spatially and temporally. Again, this leads to a much better reconstruction! Along with all this, GLOMAP has become a first-class global mapper, making it super easy to use directly within pycolmap! I'm excited to do more with this and compare it to things like pycuvslam, vipe, and other alternatives.

Pablo Vela

30,070 views • 3 months ago

More progress! I now have two Dockerized Gradio | Rerun apps. The first one takes as input a "raw" rrd file that consists of the synchronized egocentric and exocentric MP4 files. This runs the pipeline and produces an "annotated" rrd file. This has the camera parameters, 3D joints, and projected 2D joints (with 6DOF mano soon). The second app takes this "annotated" rrd file and allows for manual labeling. This is a crucial step in addressing any major failures in the pipeline. Right now, it is only the ego view that can be modified. But I'll eventually extend to all. This results in a final "gt" rrd file. From here, the plan is to improve quality and start building a data loop. Excited to start really scaling this. I'm basically going all in on keeping my data stored as Rerun rrd files. As always, I want to emphasize how crucial it is to LOOK AT YOUR data! The rrd format makes it incredibly easy to do so. Getting the data out to use is a bit of a hassle right now, but for me, it's well worth the tradeoff.
1:21

More progress! I now have two Dockerized Gradio | Rerun apps. The first one takes as input a "raw" rrd file that consists of the synchronized egocentric and exocentric MP4 files. This runs the pipeline and produces an "annotated" rrd file. This has the camera parameters, 3D joints, and projected 2D joints (with 6DOF mano soon). The second app takes this "annotated" rrd file and allows for manual labeling. This is a crucial step in addressing any major failures in the pipeline. Right now, it is only the ego view that can be modified. But I'll eventually extend to all. This results in a final "gt" rrd file. From here, the plan is to improve quality and start building a data loop. Excited to start really scaling this. I'm basically going all in on keeping my data stored as Rerun rrd files. As always, I want to emphasize how crucial it is to LOOK AT YOUR data! The rrd format makes it incredibly easy to do so. Getting the data out to use is a bit of a hassle right now, but for me, it's well worth the tradeoff.

Pablo Vela

19,527 views • 8 months ago

This is probably the biggest news yet in software going headless, and will bring knowledge work agents to the masses. The new ChatGPT agents have access to any of the tools and data you want to work with, with complete coding and tool use available to them. Here's an example of a custom sales assistant agent uses Box as a knowledge source for accessing enterprise content securely to answer questions and generate new content on the fly. The workflows can obviously vastly far more complex as the agent can use any of the tools within Box available via MCP and CLI. This precisely what agents will start to look like for knowledge work. You'll be able to spin them up in the foreground or background to help augment work. Big opportunity right now for headless platforms, and for all the new builders and designers of these agents in the enterprise.
1:06

This is probably the biggest news yet in software going headless, and will bring knowledge work agents to the masses. The new ChatGPT agents have access to any of the tools and data you want to work with, with complete coding and tool use available to them. Here's an example of a custom sales assistant agent uses Box as a knowledge source for accessing enterprise content securely to answer questions and generate new content on the fly. The workflows can obviously vastly far more complex as the agent can use any of the tools within Box available via MCP and CLI. This precisely what agents will start to look like for knowledge work. You'll be able to spin them up in the foreground or background to help augment work. Big opportunity right now for headless platforms, and for all the new builders and designers of these agents in the enterprise.

Aaron Levie

419,956 views • 2 months ago

I asked Garry Tan how to use meta prompting to get better at AI: "My partners at YC Jared Friedman and Pete Koomen showed me how to do this. You can take almost anything that you do all the time and just drop it into a context window. And then say, “Here’s a bunch of inputs and outputs." And maybe you also add a bunch of notes. And then you tell it, “Write me a prompt that can act as an agent that takes this input and makes this output over here.” You can do this for almost any type of knowledge work. And you can even introspect. "What are things you notice that I did to convert this from the input to the output?”. And then you can just start using the prompt. Initially, it’s going to suck. Because it’s just not that smart yet. But what’s funny is now, I also use it to Iterate my writing. You can be very direct, "I would never say that", "Don’t say it like this", or "Oh, you used the long word there, use the short word". Just speak to it conversationally. And then when you're happy with the output, you can use that new output to make a new prompt. "Based on this conversation, give me a better initial prompt that incorporates all the things we talked about." And you can do this with literally everything. And in theory, there’s so much it applies to that people do day-to-day. You could use it for tweets. You could use it for editing podcasts. You can use it for pretty much everything. I have a folder of prompts that I use all the time. My YouTube prompt is on v27 or something. I'll go through this process with all the different max models. I'll use GPT 5.2 Pro. I’ll use Grok. I'll use Claude. Then, I’ll take all the outputs from all the models and put them into Claude and say "Here’s my prompt, here’s the output from four LLMs, including yourself. Rate each response and tell me what the pros and cons of each approach are." And I usually say "give it to me in numbered form". And then you can agree with one, disagree with two, tell it three is this or that. And then after that, you say given all of this, synthesize it."
4:16

I asked Garry Tan how to use meta prompting to get better at AI: "My partners at YC Jared Friedman and Pete Koomen showed me how to do this. You can take almost anything that you do all the time and just drop it into a context window. And then say, “Here’s a bunch of inputs and outputs." And maybe you also add a bunch of notes. And then you tell it, “Write me a prompt that can act as an agent that takes this input and makes this output over here.” You can do this for almost any type of knowledge work. And you can even introspect. "What are things you notice that I did to convert this from the input to the output?”. And then you can just start using the prompt. Initially, it’s going to suck. Because it’s just not that smart yet. But what’s funny is now, I also use it to Iterate my writing. You can be very direct, "I would never say that", "Don’t say it like this", or "Oh, you used the long word there, use the short word". Just speak to it conversationally. And then when you're happy with the output, you can use that new output to make a new prompt. "Based on this conversation, give me a better initial prompt that incorporates all the things we talked about." And you can do this with literally everything. And in theory, there’s so much it applies to that people do day-to-day. You could use it for tweets. You could use it for editing podcasts. You can use it for pretty much everything. I have a folder of prompts that I use all the time. My YouTube prompt is on v27 or something. I'll go through this process with all the different max models. I'll use GPT 5.2 Pro. I’ll use Grok. I'll use Claude. Then, I’ll take all the outputs from all the models and put them into Claude and say "Here’s my prompt, here’s the output from four LLMs, including yourself. Rate each response and tell me what the pros and cons of each approach are." And I usually say "give it to me in numbered form". And then you can agree with one, disagree with two, tell it three is this or that. And then after that, you say given all of this, synthesize it."

The Peel

51,632 views • 4 months ago

An asian guy just found a way to use Fable 5 forever, even though Anthropic is going to make it access-limited after 12th July. He used Opus 4.8 to host the operating manual of fable 5 with just one single prompt. Access to the best model is never guaranteed. It disappeared once already this year. But you can use it forever. Here's how: 1. Extract the operating manual of Fable 5 2. Create an .md file and save it on your device 3. Create a new project "Fable 5 Brain" in Claude 4. Paste the instructions/ .md file straight into it 5. Run a pressure test to verify Save and bookmark this no matter what:
0:30

An asian guy just found a way to use Fable 5 forever, even though Anthropic is going to make it access-limited after 12th July. He used Opus 4.8 to host the operating manual of fable 5 with just one single prompt. Access to the best model is never guaranteed. It disappeared once already this year. But you can use it forever. Here's how: 1. Extract the operating manual of Fable 5 2. Create an .md file and save it on your device 3. Create a new project "Fable 5 Brain" in Claude 4. Paste the instructions/ .md file straight into it 5. Run a pressure test to verify Save and bookmark this no matter what:

Aria Westcott

359,939 views • 2 days ago

The same kinds of productivity gains we've seen in coding with AI agents are heading to the rest of knowledge work. This is the jump when you go from having a chatbot to being able to actually have an agent go off and do work for minutes or even hours and come back with a complete work output that you then review. Here's an example of the new Box Agent filling out an RFP response from an existing knowledge base. This process would normally take hours to fill out, and requires the full attention of the user doing the work. Now, you provide the Box Agent with the RFP questions, and it will go off, make a plan, extract all the relevant questions, read through existing source material to come up with an answer, and then generate a new word document as the final output. All while you're doing something else. The key to this architecture is that the agent is able to use all of the same tools in the background that a user uses to get work done. The agent can search for documents, read entire files, run scripts and tools in the background, and even be able to write code on the fly to automate tasks it hasn't seen before. And best of all, the Box Agent will (soon) work from the Box MCP and CLI so you can invoke it in any agentic system as a step in a process. This kind of agent complexity would have been impossible even 6 months ago. Models consistently failed at tracking long running tasks or using the right tools at the right moment for the task. But this is all now possible because of models like GPT-5.4, Opus 4.6, and Gemini 3, and is only getting better by the month. Just as we moved from engineers writing code and using AI as an assistant to answer questions, in many areas of knowledge work -like legal, finance, consulting, sales, marketing, and more- when we have a problem we'll just kick off the AI agent to just go work on it for us in the background.
2:14

The same kinds of productivity gains we've seen in coding with AI agents are heading to the rest of knowledge work. This is the jump when you go from having a chatbot to being able to actually have an agent go off and do work for minutes or even hours and come back with a complete work output that you then review. Here's an example of the new Box Agent filling out an RFP response from an existing knowledge base. This process would normally take hours to fill out, and requires the full attention of the user doing the work. Now, you provide the Box Agent with the RFP questions, and it will go off, make a plan, extract all the relevant questions, read through existing source material to come up with an answer, and then generate a new word document as the final output. All while you're doing something else. The key to this architecture is that the agent is able to use all of the same tools in the background that a user uses to get work done. The agent can search for documents, read entire files, run scripts and tools in the background, and even be able to write code on the fly to automate tasks it hasn't seen before. And best of all, the Box Agent will (soon) work from the Box MCP and CLI so you can invoke it in any agentic system as a step in a process. This kind of agent complexity would have been impossible even 6 months ago. Models consistently failed at tracking long running tasks or using the right tools at the right moment for the task. But this is all now possible because of models like GPT-5.4, Opus 4.6, and Gemini 3, and is only getting better by the month. Just as we moved from engineers writing code and using AI as an assistant to answer questions, in many areas of knowledge work -like legal, finance, consulting, sales, marketing, and more- when we have a problem we'll just kick off the AI agent to just go work on it for us in the background.

Aaron Levie

24,618 views • 3 months ago

I transformed my physical door into a mixed-reality portal that can bring you either to your next room or to magical virtual worlds! You can select where you want to go (like when you choose the floor in an elevator) and then when you open the door, you can exactly go to that place, physical or virtual that it may be. What I love the most is that this demo connects the physicality of the door with mixed reality, so you use a real tool to be able to access virtual realities. I've made this demo using #unity3d and #MetaQuest SDK. I've used Meta Scene Setup, so the demo is fully dynamic: it finds my door and adapts all the interfaces to it. Theoretically, it could work with all sliding doors that work like mine with no modifications to the APK. As usual, I'm thinking of making a blog post where I explain how I did it... let me know if you could be interested! And let me know also if you could be interested in some collaborations with me ;) The full video of the demo is on Youtube: #metaverse #VirtualReality #SpatialComputing
2:07

I transformed my physical door into a mixed-reality portal that can bring you either to your next room or to magical virtual worlds! You can select where you want to go (like when you choose the floor in an elevator) and then when you open the door, you can exactly go to that place, physical or virtual that it may be. What I love the most is that this demo connects the physicality of the door with mixed reality, so you use a real tool to be able to access virtual realities. I've made this demo using #unity3d and #MetaQuest SDK. I've used Meta Scene Setup, so the demo is fully dynamic: it finds my door and adapts all the interfaces to it. Theoretically, it could work with all sliding doors that work like mine with no modifications to the APK. As usual, I'm thinking of making a blog post where I explain how I did it... let me know if you could be interested! And let me know also if you could be interested in some collaborations with me ;) The full video of the demo is on Youtube: #metaverse #VirtualReality #SpatialComputing

TonyVT SkarredGhost

20,071 views • 2 years ago

SURELY THAT’S DEFAMATORY ? So that piece of work Brittany’s new boyfriend David Sharaz states that I’m the type of person that would publicly tell a rape victim to "just suck it up" - as they all laugh - and Channel 7 decides to broadcast that statement That’s nothing but a false statement that I take great offence to. And this total piece of work is saying this as he’s plotting with The Project to USE an alleged rape victim for a political hit job. When an allegation such as this is made, it should have been a matter for the police and the courts
1:14

SURELY THAT’S DEFAMATORY ? So that piece of work Brittany’s new boyfriend David Sharaz states that I’m the type of person that would publicly tell a rape victim to "just suck it up" - as they all laugh - and Channel 7 decides to broadcast that statement That’s nothing but a false statement that I take great offence to. And this total piece of work is saying this as he’s plotting with The Project to USE an alleged rape victim for a political hit job. When an allegation such as this is made, it should have been a matter for the police and the courts

Craig Kelly:🇦🇺Foundation for Economic Education

55,316 views • 3 years ago

Jacob Tierney discusses his process for writing Heated Rivalry and outlining season two: "The book [Heated Rivalry] is in five parts and very quickly I was like, part one, episode one. Part two, episode two. It was very clear to me. …So in this case, I actually did not outline. Because I was just using these parts of this book, and I knew these books so well at this point. Something that I did, and that I'm trying to do again now when I'm writing the new season, is I'm trying to use—Because there's a dreaminess to this show, I try to use my memory as a guide. I'm like, what do I remember? And then I try to give primacy to the stuff that I remember and that has stuck in my brain all these years with this story. So I’m like, oh I have to do that! And that's a nice way for me to kind of center things. Where if I have to do that, then it means maybe I don't have to do this, and it maybe means I want to combine or collapse different things. Because if this is going to take up—If one incident that I'm thinking of is going to take up the space in an episode that I think of as the heart, …then you don't need to do a first version of it in the same way, you know? Little things like that. That being said, for this season because I'm working with a co-writer as well, we have outlined everything. And every time, I do approach outlining like a teenager, where I'm like, [modulates voice] I don't want to. But then when I do it, I'm always like, why don't I always do this? It makes everything so much easier. So I kind of gaslight myself in that way." ✍🏼 transcription via Heated Rivalry News & Updates. Please credit if reposting. 🗣️ quote via q&a with Stage 32 on March 24, 2026. 🔗
2:12

Jacob Tierney discusses his process for writing Heated Rivalry and outlining season two: "The book [Heated Rivalry] is in five parts and very quickly I was like, part one, episode one. Part two, episode two. It was very clear to me. …So in this case, I actually did not outline. Because I was just using these parts of this book, and I knew these books so well at this point. Something that I did, and that I'm trying to do again now when I'm writing the new season, is I'm trying to use—Because there's a dreaminess to this show, I try to use my memory as a guide. I'm like, what do I remember? And then I try to give primacy to the stuff that I remember and that has stuck in my brain all these years with this story. So I’m like, oh I have to do that! And that's a nice way for me to kind of center things. Where if I have to do that, then it means maybe I don't have to do this, and it maybe means I want to combine or collapse different things. Because if this is going to take up—If one incident that I'm thinking of is going to take up the space in an episode that I think of as the heart, …then you don't need to do a first version of it in the same way, you know? Little things like that. That being said, for this season because I'm working with a co-writer as well, we have outlined everything. And every time, I do approach outlining like a teenager, where I'm like, [modulates voice] I don't want to. But then when I do it, I'm always like, why don't I always do this? It makes everything so much easier. So I kind of gaslight myself in that way." ✍🏼 transcription via Heated Rivalry News & Updates. Please credit if reposting. 🗣️ quote via q&a with Stage 32 on March 24, 2026. 🔗

Heated Rivalry News & Updates

60,684 views • 3 months ago

This agent relies on a marketplace of 2.8M+ specialized agents to complete a task. This is how I imagine the future of personal agents. Here is how this works: 1. Agentverse is a marketplace of agents 2. Anyone can build and publish an agent there 3. ASI:One is a personal agent with access to the marketplace 4. To solve a task, ASI:One finds the best agent for the job In other words: You have an agent that can solve pretty much *anything* because it knows how to outsource the work to a network of specialized agents. And you don't need to do any setup to get this working. The way ranking agents in the marketplace works is pretty cool: Fetch.ai, the company behind it, created an AgentRank algorithm that works much like Google's PageRank. Every time an agent calls another, it forms a connection. Over time, these connections build into a graph of who relies on whom. An agent that is frequently called by other reputable agents scores higher and surfaces higher in results. The end result here is that your agent will consistently use the best agents in the marketplace to accomplish a goal. You should give this a try: You don't need to pay or even register to try these agents. Thanks to the team for partnering with me on this post.
4:16

This agent relies on a marketplace of 2.8M+ specialized agents to complete a task. This is how I imagine the future of personal agents. Here is how this works: 1. Agentverse is a marketplace of agents 2. Anyone can build and publish an agent there 3. ASI:One is a personal agent with access to the marketplace 4. To solve a task, ASI:One finds the best agent for the job In other words: You have an agent that can solve pretty much *anything* because it knows how to outsource the work to a network of specialized agents. And you don't need to do any setup to get this working. The way ranking agents in the marketplace works is pretty cool: Fetch.ai, the company behind it, created an AgentRank algorithm that works much like Google's PageRank. Every time an agent calls another, it forms a connection. Over time, these connections build into a graph of who relies on whom. An agent that is frequently called by other reputable agents scores higher and surfaces higher in results. The end result here is that your agent will consistently use the best agents in the marketplace to accomplish a goal. You should give this a try: You don't need to pay or even register to try these agents. Thanks to the team for partnering with me on this post.

Santiago

18,665 views • 18 days ago

This is one of the wildest things I have built with AI so far 😱 Asked Claude Fable 5 to create a fully interactive 3D crime investigation board using Three.js and it did it in ONE Prompt. It took one more prompt to give it a polished UI inspired by a reference design that I created with GPT Image. What surprised me most was not how it looked. It was how it worked. Fable built a complete Verlet rope physics system with 46 particles in every thread. It handled gravity, damping, distance constraints and all other features of file upload, text and everything. The threads/ropes even lift naturally with pinned evidence instead of clipping through it. I never asked it to implement any of that. I only described the experience I wanted. Maybe that's how we should work with AI. Instead of writing long long specs, we should explain what we want as an experience and let AI do its thing. Anyway, every day Fable 5 raises the bar for what I think AI can build. More to this Live: Code:
0:49

This is one of the wildest things I have built with AI so far 😱 Asked Claude Fable 5 to create a fully interactive 3D crime investigation board using Three.js and it did it in ONE Prompt. It took one more prompt to give it a polished UI inspired by a reference design that I created with GPT Image. What surprised me most was not how it looked. It was how it worked. Fable built a complete Verlet rope physics system with 46 particles in every thread. It handled gravity, damping, distance constraints and all other features of file upload, text and everything. The threads/ropes even lift naturally with pinned evidence instead of clipping through it. I never asked it to implement any of that. I only described the experience I wanted. Maybe that's how we should work with AI. Instead of writing long long specs, we should explain what we want as an experience and let AI do its thing. Anyway, every day Fable 5 raises the bar for what I think AI can build. More to this Live: Code:

The Bugged Dev

66,836 views • 2 days ago

Google DeepMind CEO Demis Hassabis on "leaving AI in the lab for longer” (full question + answer in the video as I've seen him misquoted). Here's what he said: "For me, the best use case of AI was to improve human health and accelerate scientific discovery..." "Given how important AGI is and how transformative a technology is, maybe the most transformative one in human history, I thought it would be best to approach the sort of latter stages of building it, which we're in now, using the scientific method, very carefully, very precisely, very thoughtfully, and rigorously with all the best scientists, in my ideal world, collaborating on in CERN-like effort, on making sure each step we understood each step each as we got to the final goal of, of building AGI.... "While we're building AGI in this careful scientific way, humanity could benefit from the proceeds of that, like cures for cancer, or maybe new energy sources or new materials… “Looking at this from 20, 30 years ago when I started out on all of this, that would have been the ideal way for it to play out, in my opinion. “Now, it didn't happen like that because technology's unpredictable and in fact, it turns out that things like language were a lot easier than we were all expecting… “We were sort of playing around with that, so were the other leading labs, but of course with ChatGPT and fair play to OpenAI, they scaled it and then they put it out there. “And I think even they say it was kind of a research experiment. They didn't realize it would go so viral. And I think none of us did and we had sort of fairly equivalent systems at the time… “Now, the downside of it is, we're in this sort of ferocious commercial pressure race that everyone's sort of locked into currently. “And then on top of that, there's geopolitical issues like the US-China race and so on. So there's sort of multiple levels of pressure to sort of move fast. So the benefit of that, of course, you get faster progress, obviously. The progress is just at lightning speed these days. So that's good for all the good use cases. The second benefit is that everybody, all of the viewers out there, everyone, you're all getting to use the most cutting edge AI technology, perhaps only three to six months behind what is actually in the labs. So that's kind of mind blowing. “It's also great because I think it gives everyone a feeling for, it's democratizing AI. It's giving everyone a feeling for what it's like to interact with cutting edge AI and what it can do and what it can't do… “So I think there's positives and negatives about the way it's gone. It's not the way I dreamed about years ago where we would be sort of contemplating this philosophically and carefully considering each next step. We're not in that world. And I'm, although I'm a scientist first and foremost, I'm also a pragmatic engineer. So, we have to deal with the world as we find it and make the best of that. And we try to do that by advancing the frontier, but also trying to be as responsible as we can with doing that as we deploy these, you know, very powerful technologies, like Gemini and Alphafold.”
7:32

Google DeepMind CEO Demis Hassabis on "leaving AI in the lab for longer” (full question + answer in the video as I've seen him misquoted). Here's what he said: "For me, the best use case of AI was to improve human health and accelerate scientific discovery..." "Given how important AGI is and how transformative a technology is, maybe the most transformative one in human history, I thought it would be best to approach the sort of latter stages of building it, which we're in now, using the scientific method, very carefully, very precisely, very thoughtfully, and rigorously with all the best scientists, in my ideal world, collaborating on in CERN-like effort, on making sure each step we understood each step each as we got to the final goal of, of building AGI.... "While we're building AGI in this careful scientific way, humanity could benefit from the proceeds of that, like cures for cancer, or maybe new energy sources or new materials… “Looking at this from 20, 30 years ago when I started out on all of this, that would have been the ideal way for it to play out, in my opinion. “Now, it didn't happen like that because technology's unpredictable and in fact, it turns out that things like language were a lot easier than we were all expecting… “We were sort of playing around with that, so were the other leading labs, but of course with ChatGPT and fair play to OpenAI, they scaled it and then they put it out there. “And I think even they say it was kind of a research experiment. They didn't realize it would go so viral. And I think none of us did and we had sort of fairly equivalent systems at the time… “Now, the downside of it is, we're in this sort of ferocious commercial pressure race that everyone's sort of locked into currently. “And then on top of that, there's geopolitical issues like the US-China race and so on. So there's sort of multiple levels of pressure to sort of move fast. So the benefit of that, of course, you get faster progress, obviously. The progress is just at lightning speed these days. So that's good for all the good use cases. The second benefit is that everybody, all of the viewers out there, everyone, you're all getting to use the most cutting edge AI technology, perhaps only three to six months behind what is actually in the labs. So that's kind of mind blowing. “It's also great because I think it gives everyone a feeling for, it's democratizing AI. It's giving everyone a feeling for what it's like to interact with cutting edge AI and what it can do and what it can't do… “So I think there's positives and negatives about the way it's gone. It's not the way I dreamed about years ago where we would be sort of contemplating this philosophically and carefully considering each next step. We're not in that world. And I'm, although I'm a scientist first and foremost, I'm also a pragmatic engineer. So, we have to deal with the world as we find it and make the best of that. And we try to do that by advancing the frontier, but also trying to be as responsible as we can with doing that as we deploy these, you know, very powerful technologies, like Gemini and Alphafold.”

Cleo Abram

64,840 views • 3 months ago

I've been on a SLAM/SFM kick. It's one of the more underexplored and lacking areas when it comes to human teleop/data collections, so I've brought over Deep Patch Visual Odometry/SLAM to Rerun and Gradio. With this example, we now have 1. pycuvslam 2. pycolmap/glomap 3. mast3r-slam 4. dpvo/slam all integrated into rerun. The question becomes, which method should be used in what situations? They all make different trade-offs with different camera requirements and throughput/accuracy. What about when a new method comes out? Now that I have several different methods, I plan to use VSLAM-LAB for evaluation. It uses prefix.dev to isolate all the dependencies of each of these methods and easily compare them against each other. In particular, I'll be converting the data preprocessing, algorithm outputs, and evaluation into rerun recordings (rrd files). This will allow both programmatic querying of anything stored in the files (which method had the highest ATE-to-FPS ratio? Which dataset/sequence caused the most difficulty? etc. etc.), all with easy visual inspection using the rerun server to link them all together. Another really important side effect of this is how it impacts agents. As Karpathy said ``` LLMs are exceptionally good at looping until they meet specific goals, and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria, and watch it go. ``` by having accuracy and throughput metrics deeply tied with human inspectable artifacts. One can really accelerate agentic development with an actual understanding of how the method/data performs. I think this is another killer use case that I'll be really leaning into to make ingestion of new datasets/methods trivial with an agent. I'm making it my mission for folks to understand that rerun as a visualization tool only scratches the surface of what its true benefit is. Deep integration between data and visuals, with powerful query capabilities. I'll be focusing on the SLAM use case first and then bringing this into the full egocentric/exocentric data collection domain!
0:32

I've been on a SLAM/SFM kick. It's one of the more underexplored and lacking areas when it comes to human teleop/data collections, so I've brought over Deep Patch Visual Odometry/SLAM to Rerun and Gradio. With this example, we now have 1. pycuvslam 2. pycolmap/glomap 3. mast3r-slam 4. dpvo/slam all integrated into rerun. The question becomes, which method should be used in what situations? They all make different trade-offs with different camera requirements and throughput/accuracy. What about when a new method comes out? Now that I have several different methods, I plan to use VSLAM-LAB for evaluation. It uses prefix.dev to isolate all the dependencies of each of these methods and easily compare them against each other. In particular, I'll be converting the data preprocessing, algorithm outputs, and evaluation into rerun recordings (rrd files). This will allow both programmatic querying of anything stored in the files (which method had the highest ATE-to-FPS ratio? Which dataset/sequence caused the most difficulty? etc. etc.), all with easy visual inspection using the rerun server to link them all together. Another really important side effect of this is how it impacts agents. As Karpathy said ``` LLMs are exceptionally good at looping until they meet specific goals, and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria, and watch it go. ``` by having accuracy and throughput metrics deeply tied with human inspectable artifacts. One can really accelerate agentic development with an actual understanding of how the method/data performs. I think this is another killer use case that I'll be really leaning into to make ingestion of new datasets/methods trivial with an agent. I'm making it my mission for folks to understand that rerun as a visualization tool only scratches the surface of what its true benefit is. Deep integration between data and visuals, with powerful query capabilities. I'll be focusing on the SLAM use case first and then bringing this into the full egocentric/exocentric data collection domain!

Pablo Vela

40,864 views • 2 months ago

🇧🇷 Carlos confirming he did use the Brasil Sprint as FP2 but still it didn’t help much for learning “we didn't catch a glimpse of your race, they didn't show it” carlos: “no, there was nothing to see about my race.. tough race but we tried to used it to try things ahead of quali, i started from the pit lane, we changed a lot of things to see what conclusions we gathered from what worked and what didn't, and see if it can gives us any extra information bc this weekend it seems that aston and sauber are clearly faster than us” “it was like fp2 then?” carlos: “for us it was, there was even a moment during the race i opened a gap with the cars in front to test a little the car, honestly i couldn't get all the information i wanted but we'll find something”
1:04

🇧🇷 Carlos confirming he did use the Brasil Sprint as FP2 but still it didn’t help much for learning “we didn't catch a glimpse of your race, they didn't show it” carlos: “no, there was nothing to see about my race.. tough race but we tried to used it to try things ahead of quali, i started from the pit lane, we changed a lot of things to see what conclusions we gathered from what worked and what didn't, and see if it can gives us any extra information bc this weekend it seems that aston and sauber are clearly faster than us” “it was like fp2 then?” carlos: “for us it was, there was even a moment during the race i opened a gap with the cars in front to test a little the car, honestly i couldn't get all the information i wanted but we'll find something”

ari 🦄✨

52,940 views • 8 months ago

What does the reputation model look like for agents? (alpha leak below) And how do we associate the proofs that we have about human beings with the agents who represent them? You may have heard of a process called KYC or Know Your Customer. That's very common with traditional financial applications and services. We have introduced a concept that we call KYA or Know Your Agent, which is a structured way to be able to express what model, how data was used in training, who the deployer is, what entities this agent instance is accountable back to, providing not only provenance but identity of the associated organization or entity. That's also another root of trust that we think about a lot: Enterprises and organizations tied back to things like their domains. To share a little bit of an alpha leak here, a product that we're excited to be rolling out in the next few weeks will allow our enterprise partners to more easily verify and prove the traits and capabilities of their teams as well as their counterparties. On the agent front, that makes it really easy to prove that an agent is acting on behalf of a given business or entity. We've already seen lawsuits where the absence of such technology has been a huge risk, such as with airlines that incorporate ChatGPT wrappers in their support pages. And then those AI enabled interactions end up making up plane tickets that don't exist and those airlines have to honor them. As small of an example as that might be, being able to prove agent accountability also unlocks a huge set of opportunities for use in enterprise for those agent to agent interactions. The Deep Trust Framework that our team has put together that we're excited to be bringing into a friendly SDK form in the next few weeks for some of our partners includes those reputation based capabilities, so how you can basically keep track of the interactions an agent has had, associate all of that to the entity to which they're accountable, and then that creates a sustainable reputation model for these agent to agent Interactions. Source: Billions CEO Evin McMullen evin speaking at House of Chimera Spaces Event Dec 3, 2025
2:06

What does the reputation model look like for agents? (alpha leak below) And how do we associate the proofs that we have about human beings with the agents who represent them? You may have heard of a process called KYC or Know Your Customer. That's very common with traditional financial applications and services. We have introduced a concept that we call KYA or Know Your Agent, which is a structured way to be able to express what model, how data was used in training, who the deployer is, what entities this agent instance is accountable back to, providing not only provenance but identity of the associated organization or entity. That's also another root of trust that we think about a lot: Enterprises and organizations tied back to things like their domains. To share a little bit of an alpha leak here, a product that we're excited to be rolling out in the next few weeks will allow our enterprise partners to more easily verify and prove the traits and capabilities of their teams as well as their counterparties. On the agent front, that makes it really easy to prove that an agent is acting on behalf of a given business or entity. We've already seen lawsuits where the absence of such technology has been a huge risk, such as with airlines that incorporate ChatGPT wrappers in their support pages. And then those AI enabled interactions end up making up plane tickets that don't exist and those airlines have to honor them. As small of an example as that might be, being able to prove agent accountability also unlocks a huge set of opportunities for use in enterprise for those agent to agent interactions. The Deep Trust Framework that our team has put together that we're excited to be bringing into a friendly SDK form in the next few weeks for some of our partners includes those reputation based capabilities, so how you can basically keep track of the interactions an agent has had, associate all of that to the entity to which they're accountable, and then that creates a sustainable reputation model for these agent to agent Interactions. Source: Billions CEO Evin McMullen evin speaking at House of Chimera Spaces Event Dec 3, 2025

Billions

68,458 views • 7 months ago

Designed this site in Fable 5 with just a few prompts. And no, this is not a lame post where I attach my actual (not AI) work on an AI post to get views by making fun of a launch. This is actually Fable 5, as you can clearly see (with some of the elements that scream AI generated). But man is it doing a great job!
0:22

Designed this site in Fable 5 with just a few prompts. And no, this is not a lame post where I attach my actual (not AI) work on an AI post to get views by making fun of a launch. This is actually Fable 5, as you can clearly see (with some of the elements that scream AI generated). But man is it doing a great job!

Asaad Mahmood - The Small Square

106,914 views • 1 month ago

New short course: Build AI Apps with MCP Servers: Working with Box Files, built with Box and taught by Ben Kus , their CTO. Many AI applications require custom code for basic file operations. The Model Context Protocol (MCP) standardizes this by letting you offload file tasks to dedicated servers that provide tools an LLM can use directly. In this course, you'll process documents stored in a Box folder using the Box MCP server. Rather than writing custom integration code to connect to the Box API and download files, you'll design your application to use the tools provided via MCP. Skills you'll gain: - Build an LLM-powered document processing app, using the Box MCP server to access files - Design a multi-agent system using Google's Agent Development Kit (ADK), consisting of specialized agents for file operations - Coordinate the multi-agent workflow through an orchestrator that uses the Agent2Agent (A2A) protocol to connect to the agents You'll start with a local file-processing app, refactor it to work with Box's MCP server, then evolve it into a multi-agent system. Sign up here:
1:48

New short course: Build AI Apps with MCP Servers: Working with Box Files, built with Box and taught by Ben Kus , their CTO. Many AI applications require custom code for basic file operations. The Model Context Protocol (MCP) standardizes this by letting you offload file tasks to dedicated servers that provide tools an LLM can use directly. In this course, you'll process documents stored in a Box folder using the Box MCP server. Rather than writing custom integration code to connect to the Box API and download files, you'll design your application to use the tools provided via MCP. Skills you'll gain: - Build an LLM-powered document processing app, using the Box MCP server to access files - Design a multi-agent system using Google's Agent Development Kit (ADK), consisting of specialized agents for file operations - Coordinate the multi-agent workflow through an orchestrator that uses the Agent2Agent (A2A) protocol to connect to the agents You'll start with a local file-processing app, refactor it to work with Box's MCP server, then evolve it into a multi-agent system. Sign up here:

Andrew Ng

81,563 views • 9 months ago

American Surgeon shows the actual letter from UnitedHealthcare DENYING a patient in emergency condition from receiving care “This is a woman who was in the emergency room with pulmonary embolisms” “I think we all knew this would happen. I had another patient come in and share with me that UnitedHealthcare denied her inpatient's day. So this is a patient who had shortness of breath and some chest pain, and she just knew that something wasn't right in her body. She had a family history of blood clots and she'd had a deep flap surgery a couple of weeks ago. She went to the hospital and they saw her and they found that she had a life threatening condition known as pulmonary embolisms. So she was admitted to the hospital and taken care of really well by the doctors there. And they ordered all the right things. After a couple of days, she was discharged. She got a letter from UnitedHealthcare explaining that they didn't agree with the level of her care and that they would not cover it. So I'm gonna share some of the language of that letter with you, and I want you to know that my patient that we talked about previously who had her surgery denied had almost exactly the same letter shared. So there's some troubling things in this letter. I think this term is really interesting. United is saying they reviewed the request for inpatient admission. So let's all just pause and consider that. This is a woman who was in the emergency room with pulmonary embolisms, and the doctor wasn't really requesting anything. They were saying this patient needs to be in the hospital. But an insurance company sees this as a request, and that's part of this prior auth environment that we're living in. So I think it's important as patients and as physicians to just acknowledge that this is our reality now. Someone can think that there's a good medical decision for you and can write orders and wanna do the right thing for you, but your insurance company is seeing that as a request and deciding whether or not they wanna do it. One of the criteria that this insurance company used to decide whether or not to accept or deny this request was whether it's medically necessary. And it's so interesting that we're letting insurance companies and the doctors who work for insurance companies determine what's medically necessary and not just the doctor in front of the patient in the emergency room. So this is a really bold statement from UnitedHealthcare for my patient. They say you did not have to be admitted as an inpatient to the hospital for this care. I think we all need to just reflect on that. An insurance company is telling a patient and her doctor that they disagree with the plan of care to keep that patient safe. I know that this is boiling down to whether it's an inpatient admission or an observation admission, and that's really about money. But what I wanna point out to you is they're making medical decisions. This insurance company is actually weighing in and disagreeing with a doctor who made a medical decision to admit this patient for her safety. So this specific sentence, when a doctor or facility treats a patient above the recommended level of care, we cannot cover it. What the heck? That's what we do. We go above and beyond as physicians. It's clear that insurance companies don't, and they're actually saying it here.”
2:35

American Surgeon shows the actual letter from UnitedHealthcare DENYING a patient in emergency condition from receiving care “This is a woman who was in the emergency room with pulmonary embolisms” “I think we all knew this would happen. I had another patient come in and share with me that UnitedHealthcare denied her inpatient's day. So this is a patient who had shortness of breath and some chest pain, and she just knew that something wasn't right in her body. She had a family history of blood clots and she'd had a deep flap surgery a couple of weeks ago. She went to the hospital and they saw her and they found that she had a life threatening condition known as pulmonary embolisms. So she was admitted to the hospital and taken care of really well by the doctors there. And they ordered all the right things. After a couple of days, she was discharged. She got a letter from UnitedHealthcare explaining that they didn't agree with the level of her care and that they would not cover it. So I'm gonna share some of the language of that letter with you, and I want you to know that my patient that we talked about previously who had her surgery denied had almost exactly the same letter shared. So there's some troubling things in this letter. I think this term is really interesting. United is saying they reviewed the request for inpatient admission. So let's all just pause and consider that. This is a woman who was in the emergency room with pulmonary embolisms, and the doctor wasn't really requesting anything. They were saying this patient needs to be in the hospital. But an insurance company sees this as a request, and that's part of this prior auth environment that we're living in. So I think it's important as patients and as physicians to just acknowledge that this is our reality now. Someone can think that there's a good medical decision for you and can write orders and wanna do the right thing for you, but your insurance company is seeing that as a request and deciding whether or not they wanna do it. One of the criteria that this insurance company used to decide whether or not to accept or deny this request was whether it's medically necessary. And it's so interesting that we're letting insurance companies and the doctors who work for insurance companies determine what's medically necessary and not just the doctor in front of the patient in the emergency room. So this is a really bold statement from UnitedHealthcare for my patient. They say you did not have to be admitted as an inpatient to the hospital for this care. I think we all need to just reflect on that. An insurance company is telling a patient and her doctor that they disagree with the plan of care to keep that patient safe. I know that this is boiling down to whether it's an inpatient admission or an observation admission, and that's really about money. But what I wanna point out to you is they're making medical decisions. This insurance company is actually weighing in and disagreeing with a doctor who made a medical decision to admit this patient for her safety. So this specific sentence, when a doctor or facility treats a patient above the recommended level of care, we cannot cover it. What the heck? That's what we do. We go above and beyond as physicians. It's clear that insurance companies don't, and they're actually saying it here.”

Wall Street Apes

115,691 views • 1 year ago

An LLM with native access to find and run agents is a huge thing! This large language model (called ASI:One and developed by Fetch.ai) can do this natively when you show it a problem: • Reason about the problem • Identify live agents and tools that can help solve it • Dynamically invoke these agents and tools Honestly, this is pretty nuts! You don't need to set up anything beforehand. You don't need to write any code. Here is the best part: ASI:One has access to any agent published on Agentverse, an open directory of agents! This gives the model access to *millions* of tools it can execute dynamically when solving a task. Here is a simple example (the one in the attached video): • I went to Google Maps • I copied the latitude and longitude of a location • I asked ASI:One about relevant places around that location • ASI:One went to the marketplace • It identified and called the Google Places agent • It returned a list of museums, restaurants, and landmarks near my location Pretty awesome! You can try this now by going to You don't need to pay or register to test it out. If you want to get ideas of what's possible, check Agentverse, the agent marketplace: The model will have access to all of those tools. This is pretty awesome and opens up a whole new world of possibilities! Thanks to the team for partnering with me on this post!
4:59

An LLM with native access to find and run agents is a huge thing! This large language model (called ASI:One and developed by Fetch.ai) can do this natively when you show it a problem: • Reason about the problem • Identify live agents and tools that can help solve it • Dynamically invoke these agents and tools Honestly, this is pretty nuts! You don't need to set up anything beforehand. You don't need to write any code. Here is the best part: ASI:One has access to any agent published on Agentverse, an open directory of agents! This gives the model access to *millions* of tools it can execute dynamically when solving a task. Here is a simple example (the one in the attached video): • I went to Google Maps • I copied the latitude and longitude of a location • I asked ASI:One about relevant places around that location • ASI:One went to the marketplace • It identified and called the Google Places agent • It returned a list of museums, restaurants, and landmarks near my location Pretty awesome! You can try this now by going to You don't need to pay or register to test it out. If you want to get ideas of what's possible, check Agentverse, the agent marketplace: The model will have access to all of those tools. This is pretty awesome and opens up a whole new world of possibilities! Thanks to the team for partnering with me on this post!

Santiago

43,784 views • 1 year ago

This couple was out jogging with their new “jogging/hiking stroller” when the mother went off a small curb. As you can see in the video, the stroller did not flex for the drop and flipped over. It was advertised that it could and is labeled as a hiking-capable stroller. So, is this the manufacturer’s or company’s fault for false advertising, or should the mother not have tried to go off a curb? At what point should companies be held responsible if their product doesn’t hold up to what they’ve advertised it would!
0:31

This couple was out jogging with their new “jogging/hiking stroller” when the mother went off a small curb. As you can see in the video, the stroller did not flex for the drop and flipped over. It was advertised that it could and is labeled as a hiking-capable stroller. So, is this the manufacturer’s or company’s fault for false advertising, or should the mother not have tried to go off a curb? At what point should companies be held responsible if their product doesn’t hold up to what they’ve advertised it would!

Suzee Q

52,468 views • 4 months ago