Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

I've always believed that intelligence should come from simple, general building blocks that learn their own abstractions directly from raw data; in other words, the bitter lesson. I think we're finally seeing a real step towards AI that can learn from all data. Language modeling on raw bytes is... a microcosm of this ideal, and to me it's a central problem in AI: if we can't build LLMs that learn end-to-end from bytes, how will we handle the rest of the world's data that we understand even less? Last year we introduced hierarchical networks (H-Nets) as a first step. Our new version finds up to 3-5x efficiency gains in controlled scaling law experiments, and at larger scale matches an open-source model trained on 11T tokens with ~10x less data and compute. The posts below walk through what we found and why we think it works.show more

Albert Gu

26,902 subscribers

1,023,296 Aufrufe • vor 7 Tagen •via X (Twitter)

Wissenschaft & Technologie Bildung

Anya Rossi• Live Now

Private livecam show

16 Kommentare

Profilbild von Albert Gu

Albert Guvor 7 Tagen

Blog: Part 1 Part 2 (Voiceover created entirely by Cartesia's Sonic text-to-speech model!)

Profilbild von Nathan (Quan) Zhou

Nathan (Quan) Zhouvor 7 Tagen

This is the kind of work that makes me excited to be at @cartesia. Huge props to the team behind H-Nets~

Profilbild von Justin Waugh

Justin Waughvor 7 Tagen

So exciting to see h-net results! The In-context tokenization part is so cool - show's the real power of hnet imo. Feels like it's directionarlly a step on the path to a true continuous learner / infinite context.

Profilbild von Karan Bania

Karan Baniavor 7 Tagen

so cool to see it work on images and patchify background as one token

Profilbild von Tim Shi

Tim Shivor 7 Tagen

amazing

Profilbild von Saumya Verma

Saumya Vermavor 7 Tagen

bitter lesson never misses, we just keep forgetting it every hype cycle

Profilbild von Tommy

Tommyvor 7 Tagen

Raw bytes still force the hierarchy

Profilbild von hailports

hailportsvor 7 Tagen

raw data, huh. language already comes pre-chewed by people. the real test is what happens where nobody bothered to name things.

Profilbild von Jonathan Chang

Jonathan Changvor 7 Tagen

what happened here?

Profilbild von Charis Mitsakis

Charis Mitsakisvor 7 Tagen

the efficiency jump is huge! the bitter lesson is always true

Profilbild von Chirag

Chiragvor 7 Tagen

Other people have spoken of a cognitive core with search and tools. A lot of the initial wow came from the fact that some x million documents have been consumed and can be recalled verbatim. That's a different task from getting work done.

Profilbild von AI Quanting

AI Quantingvor 7 Tagen

What does staying at the byte level cost you at inference time?

Profilbild von Ir rev

Ir revvor 7 Tagen

does any of the insights translate to image/video-gen?

Profilbild von Gravity Research

Gravity Researchvor 7 Tagen

10x less compute is absurd

Profilbild von Nancy Xu

Nancy Xuvor 7 Tagen

Incredible work!!

Profilbild von Sophia Yang

Sophia Yangvor 7 Tagen

beautiful!

Ähnliche Videos

.L3Harris says Palantir helped them beat frontier AI models in less than 2 days, at 95% lower cost: "When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours." "The cost of our fine-tuned open source model was 95% lower than the frontier models we were using." "AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge." "We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates." " We own the model, we own the compute, we own the advantage."

.L3Harris says Palantir helped them beat frontier AI models in less than 2 days, at 95% lower cost: "When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours." "The cost of our fine-tuned open source model was 95% lower than the frontier models we were using." "AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge." "We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates." " We own the model, we own the compute, we own the advantage."

Jawwwn

1,108,745 Aufrufe • vor 1 Monat

Over the last few months, we’ve been thinking about how to learn from “off-domain” data - data from non-robot sources like video or simulation. These data sources are not quite good enough to learn policies (even monolithic VLA models) directly, but they still contain lots of information that can be useful for generalizable robot control. How can we develop robot learning models that are able to make use of this type of data for generalizable control? In new work, that we call HAMSTER, we show that VLMs can be useful for enabling robotic learning from off-domain data, but specifically when used through hierarchical VLA architectures. We show that this class of models can learn generalizable robot policies for the real world from large-scale, off-domain data. A 🧵 (1/10)

Over the last few months, we’ve been thinking about how to learn from “off-domain” data - data from non-robot sources like video or simulation. These data sources are not quite good enough to learn policies (even monolithic VLA models) directly, but they still contain lots of information that can be useful for generalizable robot control. How can we develop robot learning models that are able to make use of this type of data for generalizable control? In new work, that we call HAMSTER, we show that VLMs can be useful for enabling robotic learning from off-domain data, but specifically when used through hierarchical VLA architectures. We show that this class of models can learn generalizable robot policies for the real world from large-scale, off-domain data. A 🧵 (1/10)

Abhishek Gupta

11,994 Aufrufe • vor 1 Jahr

🤷What if we want to learn from human data... without human data? In our work NIL (No-data Imitation Learning) #CVPR2026, we explore a simple but ambitious question: Can robots learn directly from AI-generated videos without any curated demonstration data? 🔗

🤷What if we want to learn from human data... without human data? In our work NIL (No-data Imitation Learning) #CVPR2026, we explore a simple but ambitious question: Can robots learn directly from AI-generated videos without any curated demonstration data? 🔗

Chenhao Li

22,062 Aufrufe • vor 5 Monaten

David Friedberg says with open source, we no longer even need data centers for 90% of what we use AI for "I thought it was worth taking a step back to recognize where we are because in the context of 'we have to stop the super intelligence, we have to shut down the data centers,' you don't even need data centers to do 90% of what you can do with AI." "You can download it on your computer for free today to do what was the most advanced technology in human history less than a year ago. It's all here today." "And I thought it was just worth taking a beat to step back and make sure we all recognize that as we get into this debate about these, quote, labs or corporations. I don't know if it even matters anymore. I think open source is here."

David Friedberg says with open source, we no longer even need data centers for 90% of what we use AI for "I thought it was worth taking a step back to recognize where we are because in the context of 'we have to stop the super intelligence, we have to shut down the data centers,' you don't even need data centers to do 90% of what you can do with AI." "You can download it on your computer for free today to do what was the most advanced technology in human history less than a year ago. It's all here today." "And I thought it was just worth taking a beat to step back and make sure we all recognize that as we get into this debate about these, quote, labs or corporations. I don't know if it even matters anymore. I think open source is here."

dnap

27,703 Aufrufe • vor 1 Monat

#WATCH | India AI Impact Summit 2026 | Delhi: Founder Chairman and CEO of Sampark Foundation & former CEO of HCL Technologies, Vineet Nayar says, "...From an employment point of view I think it is very important for us to understand that Indian companies, including Indian IT companies, are going to be profit-driven and therefore if you believe that they are going to create employment you must be dreaming. Therefore, the question is how do we create employment in this environment, and that employment comes from mass scale startups, which is what this government has already doing. So, how do we create new sets of people who are trying to solve new sets of problems not new sets of technology and if we do that we will get it right. I think we as Indians have to be very careful on who does data belong to and that is the debate we have a problem with. The LLM models which exist worldwide are far superior than the Indian models. Unfortunately, in India, we never develop products, so therefore we do not have SLMs and LLMs which are world-class. On one side, we have global LLM products which are coming to India and trading on our Indian data. Should we allowed that or should we not allowed that? But on the other side if we don't allow that then we have the data but we don't have the LLM models. So, how do we encourage technology completely to develop the LLM models. This needs radicals strategic thinking and a very important aspect otherwise we will either give up a data. So, I think it's a very critical aspect for us to think about - who does this data belong, what is the kind of incentives we are going to give to develop LLM technologies or SLM technologies fast so that we train on our data otherwise an LLM will come in with our data and we'll immediately see return and we'll celebrate and we will do all these kind of press releases but the India will lose a competitive advantage on something which is very critical for the next decade."

#WATCH | India AI Impact Summit 2026 | Delhi: Founder Chairman and CEO of Sampark Foundation & former CEO of HCL Technologies, Vineet Nayar says, "...From an employment point of view I think it is very important for us to understand that Indian companies, including Indian IT companies, are going to be profit-driven and therefore if you believe that they are going to create employment you must be dreaming. Therefore, the question is how do we create employment in this environment, and that employment comes from mass scale startups, which is what this government has already doing. So, how do we create new sets of people who are trying to solve new sets of problems not new sets of technology and if we do that we will get it right. I think we as Indians have to be very careful on who does data belong to and that is the debate we have a problem with. The LLM models which exist worldwide are far superior than the Indian models. Unfortunately, in India, we never develop products, so therefore we do not have SLMs and LLMs which are world-class. On one side, we have global LLM products which are coming to India and trading on our Indian data. Should we allowed that or should we not allowed that? But on the other side if we don't allow that then we have the data but we don't have the LLM models. So, how do we encourage technology completely to develop the LLM models. This needs radicals strategic thinking and a very important aspect otherwise we will either give up a data. So, I think it's a very critical aspect for us to think about - who does this data belong, what is the kind of incentives we are going to give to develop LLM technologies or SLM technologies fast so that we train on our data otherwise an LLM will come in with our data and we'll immediately see return and we'll celebrate and we will do all these kind of press releases but the India will lose a competitive advantage on something which is very critical for the next decade."

ANI

18,753 Aufrufe • vor 8 Monaten

What Neuralink just achieved gives me a bizarre new idea to train AI. Below is a person controlling mouse at unreasonably high speed & precision, just by thinking. The information flow looks like screen pixels -> biological NN -> translation by an artificial NN -> continuous mouse actions. I trained lots of AI agents to play games and control browsers. It makes me wonder: What if we learn another NN to imitate the first part of the flow: pixels -> brain? Can we intercept the "chain of thought" signals that a human computes? If we use imitation learning on this data, wouldn't we be able to model something as elusive as consciousness? Is this an alternative path to learn intelligence *directly* from human inner voices, instead of indirectly from our written text tokens? Could we do it? Yes, I believe it's all technically feasible right now. Just need to scale up. Should we do it?

What Neuralink just achieved gives me a bizarre new idea to train AI. Below is a person controlling mouse at unreasonably high speed & precision, just by thinking. The information flow looks like screen pixels -> biological NN -> translation by an artificial NN -> continuous mouse actions. I trained lots of AI agents to play games and control browsers. It makes me wonder: What if we learn another NN to imitate the first part of the flow: pixels -> brain? Can we intercept the "chain of thought" signals that a human computes? If we use imitation learning on this data, wouldn't we be able to model something as elusive as consciousness? Is this an alternative path to learn intelligence directly from human inner voices, instead of indirectly from our written text tokens? Could we do it? Yes, I believe it's all technically feasible right now. Just need to scale up. Should we do it?

Jim Fan

170,541 Aufrufe • vor 2 Jahren

World Labs CEO Dr. Fei-Fei Li & SceniX Co-Founder Yunzhu Li on the data bottleneck in robotics and how world models help: "The lack of data in training, the lack of data in evaluation, this is very, very different from language models, where data is abundant on the internet." "And we know that in order for robotics to work, we have to somehow unlock the power of scaling law. But where does that come from?" "It's a profound problem that everybody's battling with in robotics." "We see a lot of unlock in being able to do this whole process through the modeling of the environments." "Being able to create these digital worlds... that's just going to unlock so much more potential for being able to replace all the costly and the unsafe data in the real environments with the data generated from the worlds for robots to be able to do scalable learning and evaluations." Fei-Fei Li Yunzhu Li martin_casado

World Labs CEO Dr. Fei-Fei Li & SceniX Co-Founder Yunzhu Li on the data bottleneck in robotics and how world models help: "The lack of data in training, the lack of data in evaluation, this is very, very different from language models, where data is abundant on the internet." "And we know that in order for robotics to work, we have to somehow unlock the power of scaling law. But where does that come from?" "It's a profound problem that everybody's battling with in robotics." "We see a lot of unlock in being able to do this whole process through the modeling of the environments." "Being able to create these digital worlds... that's just going to unlock so much more potential for being able to replace all the costly and the unsafe data in the real environments with the data generated from the worlds for robots to be able to do scalable learning and evaluations." Fei-Fei Li Yunzhu Li martin_casado

a16z

80,437 Aufrufe • vor 3 Monaten

With enough data, robots and AI can learn “world models” that let them predict the results of their actions. These models are a way to learn how embodied AI agents can perform a wide variety of useful tasks — but they require a huge amount of data. The team at General Intuition General Intuition has a solution: use data from video games! Games teach movement, problem solving, and complex spatial reasoning, and they come in a staggering diversity of forms, covering a wide variety of problems. What’s more, the captured data is high-quality, without the noise or annotation error that can come from We sat down with Pim de Witte and Adam Jelley from the General Intuition team to learn more about their history, their plans, and their philosophy.

With enough data, robots and AI can learn “world models” that let them predict the results of their actions. These models are a way to learn how embodied AI agents can perform a wide variety of useful tasks — but they require a huge amount of data. The team at General Intuition General Intuition has a solution: use data from video games! Games teach movement, problem solving, and complex spatial reasoning, and they come in a staggering diversity of forms, covering a wide variety of problems. What’s more, the captured data is high-quality, without the noise or annotation error that can come from We sat down with Pim de Witte and Adam Jelley from the General Intuition team to learn more about their history, their plans, and their philosophy.

RoboPapers

85,927 Aufrufe • vor 11 Monaten

Jason Ma reveals the first true scaling law in robotics: robots trained on 1 million hours of first-person human video predictably get better without ever seeing robot data "Yesterday we announced our new flagship robot foundation model called Dyna-2. It's very significant for the entire field for many different reasons. First of all, it's the first robot foundation model trained on at least one million hours of data. That in itself is a very challenging infrastructure challenge." "What we demonstrated is that even if the one million hours of data is entirely first-person video of humans doing manipulation tasks, we actually see scaling law for transferring to robot embodiments that the model has never seen before." "By training on more human data, we actually see predictable performance improvement on robots. That's a big deal because the biggest challenge facing robot foundation models is that we don't have enough data." "Unlike language models, there are just not enough robotics data out there, and it's very difficult to collect these robot data. Only by passively observing humans doing things can we scale this kind of data, and this is the first time we showed there is a lot of improvement that can transfer." Jason Ma Dyna Robotics

Jason Ma reveals the first true scaling law in robotics: robots trained on 1 million hours of first-person human video predictably get better without ever seeing robot data "Yesterday we announced our new flagship robot foundation model called Dyna-2. It's very significant for the entire field for many different reasons. First of all, it's the first robot foundation model trained on at least one million hours of data. That in itself is a very challenging infrastructure challenge." "What we demonstrated is that even if the one million hours of data is entirely first-person video of humans doing manipulation tasks, we actually see scaling law for transferring to robot embodiments that the model has never seen before." "By training on more human data, we actually see predictable performance improvement on robots. That's a big deal because the biggest challenge facing robot foundation models is that we don't have enough data." "Unlike language models, there are just not enough robotics data out there, and it's very difficult to collect these robot data. Only by passively observing humans doing things can we scale this kind of data, and this is the first time we showed there is a lot of improvement that can transfer." Jason Ma Dyna Robotics

MTS

18,002 Aufrufe • vor 2 Monaten

I was impressed when my team showed me this example of intelligence and behavioural generalisation. Watch this — seeing a person bow to welcome our car to move forward is definitely not something I've seen outside of Japan. Yet in this clip you can see our AI Driver recognise that this person is asking us to stop. We have the intelligence to understand the context that we shouldn't nudge around them, rather hold in place, despite no signage or red light. Then, if you keep watching, you'll see we recognise the bow and a wave as permission to drive through the road works. Less than 1% of our training data is from Japan. Our AI is trained end-to-end. There are no explicit rules or labelling, no explicit gesture or body pose recognition, this is all learned capability that is scalable and data-driven. This is how Wayve will scale autonomous driving to the world and address the long-tail of edge cases.

I was impressed when my team showed me this example of intelligence and behavioural generalisation. Watch this — seeing a person bow to welcome our car to move forward is definitely not something I've seen outside of Japan. Yet in this clip you can see our AI Driver recognise that this person is asking us to stop. We have the intelligence to understand the context that we shouldn't nudge around them, rather hold in place, despite no signage or red light. Then, if you keep watching, you'll see we recognise the bow and a wave as permission to drive through the road works. Less than 1% of our training data is from Japan. Our AI is trained end-to-end. There are no explicit rules or labelling, no explicit gesture or body pose recognition, this is all learned capability that is scalable and data-driven. This is how Wayve will scale autonomous driving to the world and address the long-tail of edge cases.

Alex Kendall

68,196 Aufrufe • vor 11 Monaten

Introducing GEN-1.5, a one-shot learner. It can learn new tasks in a few seconds. Show it what to do, and it generalizes. This capability emerged from pretraining on physical data at scale, as a step towards our mission of building general intelligence for the physical world.

Introducing GEN-1.5, a one-shot learner. It can learn new tasks in a few seconds. Show it what to do, and it generalizes. This capability emerged from pretraining on physical data at scale, as a step towards our mission of building general intelligence for the physical world.

Generalist

3,446,526 Aufrufe • vor 2 Monaten

Twelve years ago we asked the question: how can we put public data "on wheels"? Datawheel was born from vision that data alone doesn't get far. It needs vehicles to reach hearts and minds. Vehicles that make data easy to find, explore, and understand. Today, Datawheel turns twelve. In these twelve years, we have built dozens of solutions that are helping define the state of the art in public data distribution. Today we celebrate this journey. A journey with a great team and amazing clients. The journey to make public data all it can be. are some of the tools defining all that public data can be. Thank you!!!

Twelve years ago we asked the question: how can we put public data "on wheels"? Datawheel was born from vision that data alone doesn't get far. It needs vehicles to reach hearts and minds. Vehicles that make data easy to find, explore, and understand. Today, Datawheel turns twelve. In these twelve years, we have built dozens of solutions that are helping define the state of the art in public data distribution. Today we celebrate this journey. A journey with a great team and amazing clients. The journey to make public data all it can be. are some of the tools defining all that public data can be. Thank you!!!

César A. Hidalgo

12,933 Aufrufe • vor 1 Jahr

Finland President Alexander Stubb: Well, we have to be more independent in some cases. If you, for example, look at storage of personal data, obviously a lot of Europeans are angry that their personal data is stored outside of the EU. So we have to build our own data centers and our own AI factories so we can actually own the data of our own citizens. But in the end, there is no such thing as European full autonomy. You will always need to collaborate with our friends in Canada, in the US, in other partnerships around the world. ASML and other tech companies can play a crucial role both in the US and here.

Finland President Alexander Stubb: Well, we have to be more independent in some cases. If you, for example, look at storage of personal data, obviously a lot of Europeans are angry that their personal data is stored outside of the EU. So we have to build our own data centers and our own AI factories so we can actually own the data of our own citizens. But in the end, there is no such thing as European full autonomy. You will always need to collaborate with our friends in Canada, in the US, in other partnerships around the world. ASML and other tech companies can play a crucial role both in the US and here.

Clash Report

22,413 Aufrufe • vor 4 Monaten

"Claude Code is the first AI tool that can actually recreate how a data scientist thinks." Here's my new episode with Sumeet Marwaha (Head of Data at Brex) on exactly how to build an AI analyst with Claude Code. We covered: ✅ 3 queries to build a data analyst MCP ✅ How to give the analyst both data + context ✅ Exclusive Brex data on which AI tools are crushing it in startups and enterprise Some quotes from Sumeet: "I've set up dashboards many times in my career. They end up getting ignored. Claude will always read it. Claude will always ask the right questions." "You have to set up Claude Code to not just look at the data. It should be able to search Slack, find the incident, and understand why your metrics look wrong." "This tool is absolutely crushing it based on our data. It's both the startup AND enterprise coding tool of choice." 📌 Watch now:

"Claude Code is the first AI tool that can actually recreate how a data scientist thinks." Here's my new episode with Sumeet Marwaha (Head of Data at Brex) on exactly how to build an AI analyst with Claude Code. We covered: ✅ 3 queries to build a data analyst MCP ✅ How to give the analyst both data + context ✅ Exclusive Brex data on which AI tools are crushing it in startups and enterprise Some quotes from Sumeet: "I've set up dashboards many times in my career. They end up getting ignored. Claude will always read it. Claude will always ask the right questions." "You have to set up Claude Code to not just look at the data. It should be able to search Slack, find the incident, and understand why your metrics look wrong." "This tool is absolutely crushing it based on our data. It's both the startup AND enterprise coding tool of choice." 📌 Watch now:

Peter Yang

39,600 Aufrufe • vor 9 Monaten

AI only exists today because it is trained on real data. The better the data input, The better the AI output. We capture incredibly high-quality raw data of real humans, in 3D, from all angles. You do the Math ✍🏻 Our data can vastly improve and expand the capabilities of AI

AI only exists today because it is trained on real data. The better the data input, The better the AI output. We capture incredibly high-quality raw data of real humans, in 3D, from all angles. You do the Math ✍🏻 Our data can vastly improve and expand the capabilities of AI

Metahero.io

24,545 Aufrufe • vor 2 Jahren

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

Genesis AI

2,750,263 Aufrufe • vor 6 Monaten

Today we're launching a new era for Octolens. When we started Octolens, we set out to build a social listening dashboard. Today, our best customers barely open it. Daily active users are down, but revenue is up 5x over the last year. So we rebuilt Octolens as what it actually is: a data layer that plugs in wherever you work. Today we cover the entire web + our MCP and API reads and writes. In 2026, if an agent can't reach the data, it's not a data source. This is social listening for the agent era.

Today we're launching a new era for Octolens. When we started Octolens, we set out to build a social listening dashboard. Today, our best customers barely open it. Daily active users are down, but revenue is up 5x over the last year. So we rebuilt Octolens as what it actually is: a data layer that plugs in wherever you work. Today we cover the entire web + our MCP and API reads and writes. In 2026, if an agent can't reach the data, it's not a data source. This is social listening for the agent era.

Charlotte Schmitt

19,994 Aufrufe • vor 4 Monaten

Did you know Venice is building out its own data center? Venice is vertically integrating, to reduce token costs and increase its margins. From the last $VVV call with Jon Venice: "[we are] building out this initial data center and then looking even at multiple potential for additional data centers after that. I believe that the first data center is supposed to come online hopefully before the end of October [or early November] there might be a world where we decide that, you know, our margins are good enough that we can increase the amount [of $VVV burn]"

Did you know Venice is building out its own data center? Venice is vertically integrating, to reduce token costs and increase its margins. From the last $VVV call with Jon Venice: "[we are] building out this initial data center and then looking even at multiple potential for additional data centers after that. I believe that the first data center is supposed to come online hopefully before the end of October [or early November] there might be a world where we decide that, you know, our margins are good enough that we can increase the amount [of $VVV burn]"

David Hoffman

32,027 Aufrufe • vor 21 Tagen