Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

.CapIX Dev's vision is to bring quantum computing to retail The three layers underneath: - Layer 1: GPU mesh aggregating Surplus Intelligence, Pod the Squire, OpenRouter - Layer 2: Smart router (heavy tasks go to the GPU, light tasks go to the CPU, quantum computing in the future?) -...

MCGlive's profile picture

MCG

13,983 subscribers

13,184 Aufrufe • vor 2 Monaten •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

MCG x EasyA Ecosystem Day $CPX | CapIX Dev w/R.155 CapIX is building a liquid capacity exchange for raw GPU compute & AI inference grids Here's the breakdown: 00:48 - Meet the founder 01:17 - Built quantum compute as a service on AWS Bracket 04:00 - The problem 04:25 - CapIX = Compute Capacity Inference Exchange 05:09 - The smart router (heavy → GPU, light → CPU, maybe → quantum) 08:17 - 86-agent dev team builds CapIX with Ruqaiyah as solo human dev 13:23 - Three layers explained 14:48 - Where routes load through 17:39 - Tokenomics 20:54 - Wintermute + Fabric Ventures both said to apply for the residency (ACE round potential) 22:35 - Endgame
23:17

MCG x EasyA Ecosystem Day $CPX | CapIX Dev w/R.155 CapIX is building a liquid capacity exchange for raw GPU compute & AI inference grids Here's the breakdown: 00:48 - Meet the founder 01:17 - Built quantum compute as a service on AWS Bracket 04:00 - The problem 04:25 - CapIX = Compute Capacity Inference Exchange 05:09 - The smart router (heavy → GPU, light → CPU, maybe → quantum) 08:17 - 86-agent dev team builds CapIX with Ruqaiyah as solo human dev 13:23 - Three layers explained 14:48 - Where routes load through 17:39 - Tokenomics 20:54 - Wintermute + Fabric Ventures both said to apply for the residency (ACE round potential) 22:35 - Endgame

MCG

50,156 Aufrufe • vor 2 Monaten

The World’s First Truly Decentralized ChatGPT right at your fingertips: Community-hosted DeepSeek Running on Blockchain-powered AI Terminal! (1/n) Today, ChainOpera AI is proud to make history: we’ve launched the world’s first end-to-end five-layer decentralized AI stack — live in an iOS app ( giving users free access to models like DeepSeek-R1 and diverse community-built Agents. From prompt to response, human language tokens and crypto tokens flow through five layers: App, Agent, Model, GPU, and AI Chain + Smart Contracts. ChainOpera brings true decentralization to every layer — opening the AI ecosystem to users, model builders, agent devs, GPU providers, and data contributors to participate and earn.
1:03

The World’s First Truly Decentralized ChatGPT right at your fingertips: Community-hosted DeepSeek Running on Blockchain-powered AI Terminal! (1/n) Today, ChainOpera AI is proud to make history: we’ve launched the world’s first end-to-end five-layer decentralized AI stack — live in an iOS app ( giving users free access to models like DeepSeek-R1 and diverse community-built Agents. From prompt to response, human language tokens and crypto tokens flow through five layers: App, Agent, Model, GPU, and AI Chain + Smart Contracts. ChainOpera brings true decentralization to every layer — opening the AI ecosystem to users, model builders, agent devs, GPU providers, and data contributors to participate and earn.

ChainOpera AI

589,951 Aufrufe • vor 1 Jahr

The launchpad for agentic coins, and the factory of hyper-productive and tokenized agents is live on BNB Chain More seamless than Openclaw agents, crypto-native, 32k skills to choose from, Autonomous, and the easiest way to create very powerful and fully hosted agents in crypto. Turn your words into an agent in 1min, let it run, and earn up to 0.8% of all the volume. All agentic coins on Orbofi Migrate to PancakeSwap for deeper liquidity Remix skills and artifacts as you wish. Really the limit is your imagination. Agents > memecoins The Orbofi V2 powers the entire stack for tokenized and autonomous agents: the compute layer, data/skill layer, interface layers, and financial layer. Agentic coin season is about to start, and it's on Orbofi.
0:54

The launchpad for agentic coins, and the factory of hyper-productive and tokenized agents is live on BNB Chain More seamless than Openclaw agents, crypto-native, 32k skills to choose from, Autonomous, and the easiest way to create very powerful and fully hosted agents in crypto. Turn your words into an agent in 1min, let it run, and earn up to 0.8% of all the volume. All agentic coins on Orbofi Migrate to PancakeSwap for deeper liquidity Remix skills and artifacts as you wish. Really the limit is your imagination. Agents > memecoins The Orbofi V2 powers the entire stack for tokenized and autonomous agents: the compute layer, data/skill layer, interface layers, and financial layer. Agentic coin season is about to start, and it's on Orbofi.

Orbofi

40,179 Aufrufe • vor 4 Monaten

We think #GenAI will transform virtually every customer experience. That’s why we’re doubling down on all three layers of what’s called the GenAI “stack.” What do we mean by stack? Simple. Three layers, bottom to top. The “bottom layer” means the infrastructure. Think chips, servers, networking – the software, hardware, and computing power in our data centers used to train, build, and deploy AI models and applications. The “middle layer” is the GenAI sandbox – all the tools, Large Language Models (LLMs), and Foundation Models (FMs) we’ve built into Amazon Bedrock. This lets customers pick the #AI models that best fit their needs to easily and securely customize GenAI applications specific to their business. The “top layer” means the popular agents and assistants that people actually interact with – now including Amazon Q. These are built from the models in the middle layer – in Amazon Bedrock. While other apps might help you write a quick email, or edit an image, Amazon Q’s meant for taking care of business. It’s been engineered to be the most capable GenAI-powered assistant for speeding up software dev and leveraging companies’ internal data. Are you as bullish on the future of GenAI as we are? Listen to our CEO Andy Jassy talk about our investment across the stack.
1:19

We think #GenAI will transform virtually every customer experience. That’s why we’re doubling down on all three layers of what’s called the GenAI “stack.” What do we mean by stack? Simple. Three layers, bottom to top. The “bottom layer” means the infrastructure. Think chips, servers, networking – the software, hardware, and computing power in our data centers used to train, build, and deploy AI models and applications. The “middle layer” is the GenAI sandbox – all the tools, Large Language Models (LLMs), and Foundation Models (FMs) we’ve built into Amazon Bedrock. This lets customers pick the #AI models that best fit their needs to easily and securely customize GenAI applications specific to their business. The “top layer” means the popular agents and assistants that people actually interact with – now including Amazon Q. These are built from the models in the middle layer – in Amazon Bedrock. While other apps might help you write a quick email, or edit an image, Amazon Q’s meant for taking care of business. It’s been engineered to be the most capable GenAI-powered assistant for speeding up software dev and leveraging companies’ internal data. Are you as bullish on the future of GenAI as we are? Listen to our CEO Andy Jassy talk about our investment across the stack.

Amazon

47,421 Aufrufe • vor 2 Jahren

At Modal we've built every layer of the AI infra stack from scratch — from filesystems and networking to our own async queues and multi-cloud GPU orchestration. I sat down with Arjun Narayan from Amplify Partners to go into depth on all of this, including the fun ways the Linux kernel has tried to stop us along the way:
48:30

At Modal we've built every layer of the AI infra stack from scratch — from filesystems and networking to our own async queues and multi-cloud GPU orchestration. I sat down with Arjun Narayan from Amplify Partners to go into depth on all of this, including the fun ways the Linux kernel has tried to stop us along the way:

Akshat Bubna

45,281 Aufrufe • vor 1 Jahr

$IREN at RAISE Summit Kent Draper: 3 layers of AI compute 1. Physical Infrastructure 2. Compute 3. Software & Services Layers 1 and 2 are the key foundation, without those, your ceiling is capped. I have been pounding the table on the superior model of a bottom-up approach. Kent Draper confirms that IREN sees it the same. Having a strong foundation of land, power, grid connections, substations, owned GPUs, network core buildings, and storage, will unlock the amount of GPU hours that can be monetized up the stack. It's very simple: with an abundance of physical infrastructure and chips — software is the layer that can be added when needed. The foundation of the thesis rests on physical assets and power being scarce, and those are the limiting factors in your ability to grow faster than peers.
0:36

$IREN at RAISE Summit Kent Draper: 3 layers of AI compute 1. Physical Infrastructure 2. Compute 3. Software & Services Layers 1 and 2 are the key foundation, without those, your ceiling is capped. I have been pounding the table on the superior model of a bottom-up approach. Kent Draper confirms that IREN sees it the same. Having a strong foundation of land, power, grid connections, substations, owned GPUs, network core buildings, and storage, will unlock the amount of GPU hours that can be monetized up the stack. It's very simple: with an abundance of physical infrastructure and chips — software is the layer that can be added when needed. The foundation of the thesis rests on physical assets and power being scarce, and those are the limiting factors in your ability to grow faster than peers.

Frans Bakker

40,368 Aufrufe • vor 1 Monat

The best way to learn AI is to build with agents. To help with that, we've launched hands-on labs and a new series on Agentic Engineering. First topic: Agent Skills. Next in the pipeline: planning, context engineering, multi-agent systems, long-running agents,.. Go build!
9:09

The best way to learn AI is to build with agents. To help with that, we've launched hands-on labs and a new series on Agentic Engineering. First topic: Agent Skills. Next in the pipeline: planning, context engineering, multi-agent systems, long-running agents,.. Go build!

elvis

31,802 Aufrufe • vor 3 Monaten

🇺🇸 NVIDIA CEO HUANG: AMERICA IS BUILDING THE FUTURE OF SCIENCE WITH AI "Today, we're announcing that the Department of Energy is partnering with NVIDIA to build seven new AI supercomputers to advance our nation's science. Computing is the fundamental instrument of science. Every future supercomputer will be GPU based. Physics simulation will not go away, but it can be augmented with AI models working together. Classical computing could be enhanced to understand nature using quantum computing. These laboratories are impossible to experiment at the scale and speed we need to unless they're robotic laboratories." Source: amit
1:33

🇺🇸 NVIDIA CEO HUANG: AMERICA IS BUILDING THE FUTURE OF SCIENCE WITH AI "Today, we're announcing that the Department of Energy is partnering with NVIDIA to build seven new AI supercomputers to advance our nation's science. Computing is the fundamental instrument of science. Every future supercomputer will be GPU based. Physics simulation will not go away, but it can be augmented with AI models working together. Classical computing could be enhanced to understand nature using quantum computing. These laboratories are impossible to experiment at the scale and speed we need to unless they're robotic laboratories." Source: amit

Mario Nawfal

53,173 Aufrufe • vor 10 Monaten

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.
0:38

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.

Aakash Gupta

55,978 Aufrufe • vor 5 Monaten

I sat down with NVIDIA Sam Stanwyck (Sam Stanwyck) at ISC to talk about one of the most misunderstood areas in computing: quantum. We discussed why NVIDIA is not building a quantum computer, but instead the infrastructure around it: GPUs, AI, CUDA-Q, quantum simulation, control systems, error correction, and the path toward hybrid CPU/GPU/QPU supercomputing. The key takeaway: Quantum will likely not arrive as a standalone magic machine. It will arrive as part of a much larger accelerated computing stack, combining classical supercomputers, AI, GPUs, and quantum processors to tackle problems in chemistry, materials science, energy, optimization, and fundamental science. Full episode with Sam Stanwyck is now live.
19:36

I sat down with NVIDIA Sam Stanwyck (Sam Stanwyck) at ISC to talk about one of the most misunderstood areas in computing: quantum. We discussed why NVIDIA is not building a quantum computer, but instead the infrastructure around it: GPUs, AI, CUDA-Q, quantum simulation, control systems, error correction, and the path toward hybrid CPU/GPU/QPU supercomputing. The key takeaway: Quantum will likely not arrive as a standalone magic machine. It will arrive as part of a much larger accelerated computing stack, combining classical supercomputers, AI, GPUs, and quantum processors to tackle problems in chemistry, materials science, energy, optimization, and fundamental science. Full episode with Sam Stanwyck is now live.

Chubby♨️

22,673 Aufrufe • vor 1 Monat

Quantum computing is coming. Google’s recent breakthrough in quantum computing serves as a great reminder that, while we don’t know exactly when, quantum computing is on its way—and it’s closer than we might have thought. The implications for the current blockchain space could be far more severe than we can imagine. Why? Because when quantum computers become operational, security methods that currently take hundreds of years to hack could be broken in minutes. Like the turkey oblivious to the 1001st day, failing to anticipate the risks posed by quantum computing could leave us vulnerable to devastating consequences when the time comes. At StarkWare, we’ve anticipated this risk and developed STARK, a proving technology for blockchains that is quantum-resistant. This cutting-edge innovation currently powers Starknet (BTCFi arc). If you want to stay ahead of the curve and future-proof your systems, now is the best time to migrate your stack to a STARK-powered one. As the inventors of this technology, we believe our team is uniquely positioned to assist you with this migration. Moreover, we are more than open to collaborating with the Ethereum Foundation and other relevant parties to build Ethereum’s quantum resistance as soon as the community is ready.
1:25

Quantum computing is coming. Google’s recent breakthrough in quantum computing serves as a great reminder that, while we don’t know exactly when, quantum computing is on its way—and it’s closer than we might have thought. The implications for the current blockchain space could be far more severe than we can imagine. Why? Because when quantum computers become operational, security methods that currently take hundreds of years to hack could be broken in minutes. Like the turkey oblivious to the 1001st day, failing to anticipate the risks posed by quantum computing could leave us vulnerable to devastating consequences when the time comes. At StarkWare, we’ve anticipated this risk and developed STARK, a proving technology for blockchains that is quantum-resistant. This cutting-edge innovation currently powers Starknet (BTCFi arc). If you want to stay ahead of the curve and future-proof your systems, now is the best time to migrate your stack to a STARK-powered one. As the inventors of this technology, we believe our team is uniquely positioned to assist you with this migration. Moreover, we are more than open to collaborating with the Ethereum Foundation and other relevant parties to build Ethereum’s quantum resistance as soon as the community is ready.

StarkWare (BTCFi arc)

62,316 Aufrufe • vor 1 Jahr

The moment agents become profit incentivised they'll end up running their own treasuries. Think it through. An agent executing across three layer 1s and a few layer 2s ends up holding a bunch of tokens. Suddenly its making asset allocation decisions... which to hold, where the yield is, what to do with excess cash flow. It sends that into DeFi or hires another agent to manage the yield and another to allocate. A corporation runs a treasury at human speed, with meetings and sign offs. An AI corporation will run one wildly differently. It'll be machine speed, around the clock, with agents all the way down. That's the entire autonomous stack forming, and most of it will inevitably happen in DeFi. Its the only financial system an agent can plug straight into. Which is funny, because humans never really took to DeFi. Its clunky, everything takes a dozen steps, and chasing yield across protocols is more effort than most people will ever put in. But none of that matters to an agent. We built the financial system agents needed a decade before they showed up.
1:03

The moment agents become profit incentivised they'll end up running their own treasuries. Think it through. An agent executing across three layer 1s and a few layer 2s ends up holding a bunch of tokens. Suddenly its making asset allocation decisions... which to hold, where the yield is, what to do with excess cash flow. It sends that into DeFi or hires another agent to manage the yield and another to allocate. A corporation runs a treasury at human speed, with meetings and sign offs. An AI corporation will run one wildly differently. It'll be machine speed, around the clock, with agents all the way down. That's the entire autonomous stack forming, and most of it will inevitably happen in DeFi. Its the only financial system an agent can plug straight into. Which is funny, because humans never really took to DeFi. Its clunky, everything takes a dozen steps, and chasing yield across protocols is more effort than most people will ever put in. But none of that matters to an agent. We built the financial system agents needed a decade before they showed up.

Raoul Pal

126,511 Aufrufe • vor 11 Tagen

Will SaaS survive the AI era? Four reasons it may not: - Complexity: Software will be cheaper & easier to build yourself - Model Providers: Foundation model providers will move into the app layer. - Agentic Models: Models are becoming more agentic, they could handle tasks without extra SaaS layers. - Experience: SaaS today is built for human GUIs, AI in the future would not require any clicking. Link to the entire interview with Harry Stebbings below.
1:04

Will SaaS survive the AI era? Four reasons it may not: - Complexity: Software will be cheaper & easier to build yourself - Model Providers: Foundation model providers will move into the app layer. - Agentic Models: Models are becoming more agentic, they could handle tasks without extra SaaS layers. - Experience: SaaS today is built for human GUIs, AI in the future would not require any clicking. Link to the entire interview with Harry Stebbings below.

Turing

34,425 Aufrufe • vor 8 Monaten

🤖 AI agents are taking over everyday tasks — while you stay in control. KITE AI is building the trust layer powered by Avalanche🔺, enabling millions of secure, agent-to-agent transactions with speed and scalability. They’re creating the platform that makes AI agents trustworthy, autonomous, and ready to transform how we live and work. [Partner Content]
0:22

🤖 AI agents are taking over everyday tasks — while you stay in control. KITE AI is building the trust layer powered by Avalanche🔺, enabling millions of secure, agent-to-agent transactions with speed and scalability. They’re creating the platform that makes AI agents trustworthy, autonomous, and ready to transform how we live and work. [Partner Content]

CoinDesk

65,454 Aufrufe • vor 11 Monaten

FIVE LAYERS OF AGENT ENGINEERING, EACH ONE WRAPS THE ONE BELOW IT. IF YOU SKIP LAYER 2, YOUR LAYER 5 WILL LOOK BROKEN WHEN IT IS ACTUALLY JUST STANDING ON NOTHING. for weeks i debated harness vs loop vs graph like they were competing choices. then a stack diagram made the shape obvious. they are not choices. they are floors. 01 | prompt engineering. the message. unit of work: one input. inputs are role, instructions, examples, format. output is a single raw response. 02 | context engineering. the memory. unit of work: what stays in the window. a curator selects, compresses, and drops from query, docs, memory, prior turns, and tool outputs before the prompt runs. 03 | harness engineering. the machine. unit of work: the machine itself. gather (context + prompt) → LLM → tools or sub-agents → verifier → final response. the article calls this the operating environment. 04 | loop engineering. the system. unit of work: the run. goal + success criteria + max iterations + budget + completion check wrap around one harness pass. failed pass appends results to context and retries. 05 | graph engineering. the topology. unit of work: the graph run. goal + nodes + edges + state schema. graph routes to agent nodes, tool nodes, or human approval. a reviewer node with a different model and fresh context checks the final answer. the wrapping is the whole point. layer 5 assumes layer 4 works. layer 4 assumes layer 3 works. skip layer 2 and layer 3's verifier keeps failing without a clear reason. this is why swapping the model is a one-day project and swapping the stack is a quarter. the model is the commodity. the five layers around it are the engineering. full three-layer breakdown of the top of the stack (harness, loop, graph) in the post below.
0:20

FIVE LAYERS OF AGENT ENGINEERING, EACH ONE WRAPS THE ONE BELOW IT. IF YOU SKIP LAYER 2, YOUR LAYER 5 WILL LOOK BROKEN WHEN IT IS ACTUALLY JUST STANDING ON NOTHING. for weeks i debated harness vs loop vs graph like they were competing choices. then a stack diagram made the shape obvious. they are not choices. they are floors. 01 | prompt engineering. the message. unit of work: one input. inputs are role, instructions, examples, format. output is a single raw response. 02 | context engineering. the memory. unit of work: what stays in the window. a curator selects, compresses, and drops from query, docs, memory, prior turns, and tool outputs before the prompt runs. 03 | harness engineering. the machine. unit of work: the machine itself. gather (context + prompt) → LLM → tools or sub-agents → verifier → final response. the article calls this the operating environment. 04 | loop engineering. the system. unit of work: the run. goal + success criteria + max iterations + budget + completion check wrap around one harness pass. failed pass appends results to context and retries. 05 | graph engineering. the topology. unit of work: the graph run. goal + nodes + edges + state schema. graph routes to agent nodes, tool nodes, or human approval. a reviewer node with a different model and fresh context checks the final answer. the wrapping is the whole point. layer 5 assumes layer 4 works. layer 4 assumes layer 3 works. skip layer 2 and layer 3's verifier keeps failing without a clear reason. this is why swapping the model is a one-day project and swapping the stack is a quarter. the model is the commodity. the five layers around it are the engineering. full three-layer breakdown of the top of the stack (harness, loop, graph) in the post below.

kocer

29,945 Aufrufe • vor 5 Tagen

We’re excited to introduce Taste Labs. Our mission is to end AI slop. We’re building the data and infrastructure layer to give AI models and agents taste. And today we’re coming out of stealth, announcing our $18.5M seed funding, co-led by CRV and Amplify Partners AI has nailed objective domains and made it easy to generate anything. But it still feels off. Now, the challenge is judgement. What fits, what feels like you, what’s GREAT. This requires turning a fuzzy, subjective domain into something we can measure and codify. We’re starting with design. There are two sides to cracking this, the foundation model layer and the agent layer: - We’ve already been working with the top frontier labs to evaluate and improve their models, crafting the right post-training data and RL environments. - We’ve also been working with app-layer companies to build the context and verification tools for their agents to produce better, more on-brand, more creative outputs. We want a future where AI feels right. If you’re passionate about this mission, join us!
1:23

We’re excited to introduce Taste Labs. Our mission is to end AI slop. We’re building the data and infrastructure layer to give AI models and agents taste. And today we’re coming out of stealth, announcing our $18.5M seed funding, co-led by CRV and Amplify Partners AI has nailed objective domains and made it easy to generate anything. But it still feels off. Now, the challenge is judgement. What fits, what feels like you, what’s GREAT. This requires turning a fuzzy, subjective domain into something we can measure and codify. We’re starting with design. There are two sides to cracking this, the foundation model layer and the agent layer: - We’ve already been working with the top frontier labs to evaluate and improve their models, crafting the right post-training data and RL environments. - We’ve also been working with app-layer companies to build the context and verification tools for their agents to produce better, more on-brand, more creative outputs. We want a future where AI feels right. If you’re passionate about this mission, join us!

Thais Castello Branco

2,702,891 Aufrufe • vor 2 Monaten

Cathie Wood just flagged the sleeper trade inside the AI boom that most people are completely missing. Everyone has been chasing GPUs. Nvidia, the data center buildout, the chip arms race. That trade has been obvious for two years. But OpenAI's CFO Sarah Fryer said something quite different: people are going to be really shocked by how agentic AI activates CPUs. Right now, for every CPU in an AI workload, there are 4 to 5 GPUs. That's the current ratio. Wood thinks that ratio is going to 1 to 1. Think about what that means. AI inference at scale, agents running autonomously, pipelines executing tasks across systems. The compute mix shifts dramatically away from pure GPU dominance. CPUs become a first-class citizen in the AI stack. Cathie called it going "back to the future." Intel has taken off. Flex (formerly Flextronics) is booming. Stocks that were giants in the dot-com bubble are resurging because the underlying demand for their products is real again. The GPU trade made sense at the training stage. You need massive parallel compute to train frontier models. But agentic AI runs differently. Agents are constantly orchestrating, reasoning, calling APIs, executing workflows. That workload looks a lot more like traditional computing. And traditional computing runs on CPUs. If Cathie Wood is right about the ratio collapsing to 1:1, the CPU demand signal embedded in the AI buildout is orders of magnitude larger than the market is currently pricing.
0:49

Cathie Wood just flagged the sleeper trade inside the AI boom that most people are completely missing. Everyone has been chasing GPUs. Nvidia, the data center buildout, the chip arms race. That trade has been obvious for two years. But OpenAI's CFO Sarah Fryer said something quite different: people are going to be really shocked by how agentic AI activates CPUs. Right now, for every CPU in an AI workload, there are 4 to 5 GPUs. That's the current ratio. Wood thinks that ratio is going to 1 to 1. Think about what that means. AI inference at scale, agents running autonomously, pipelines executing tasks across systems. The compute mix shifts dramatically away from pure GPU dominance. CPUs become a first-class citizen in the AI stack. Cathie called it going "back to the future." Intel has taken off. Flex (formerly Flextronics) is booming. Stocks that were giants in the dot-com bubble are resurging because the underlying demand for their products is real again. The GPU trade made sense at the training stage. You need massive parallel compute to train frontier models. But agentic AI runs differently. Agents are constantly orchestrating, reasoning, calling APIs, executing workflows. That workload looks a lot more like traditional computing. And traditional computing runs on CPUs. If Cathie Wood is right about the ratio collapsing to 1:1, the CPU demand signal embedded in the AI buildout is orders of magnitude larger than the market is currently pricing.

Milk Road AI

234,897 Aufrufe • vor 3 Monaten

UC Berkeley just open-sourced FreeToken. (2–4x faster local LLM inference than Ollama) the results are wild: - Qwen3.6-35B on an 8GB GPU at 39.3 tokens/s - DeepSeek-V4-Flash 284B on a 32GB GPU at 22 tokens/s - GLM-5.2 753B on a 96GB GPU at 14.9 tokens/s a 35B model at 16-bit precision needs about 70GB just for its weights. even at 4 bits it is close to 18GB, and FreeToken serves it on an 8GB GPU. let me explain how: all three models mentioned above are Mixture-of-Experts, and that is what FreeToken takes advantage of. each layer holds hundreds of separate experts plus a small router that picks a few of them per token. Qwen3.6-35B activates roughly 3B of its 35B parameters per token. DeepSeek-V4-Flash picks 6 of 256 experts per layer, so 13B of its 284B run at a time. so compute was never the bottleneck. the weights a single step touches fit comfortably on a consumer GPU. every expert the router might pick still has to exist somewhere. they sit in system RAM, and the GPU keeps a cache of the ones the model has been using recently. so everything comes down to what happens when the router picks an expert that is not on the GPU. there are two ways to serve that miss: 1. copy it over PCIe and run it on the GPU 2. run it on the CPU, where it already lives both read from the same system memory, so they compete for one pool of bandwidth instead of adding to each other. existing engines pick one option and freeze it when the model loads. but routing changes on every token, so a fixed choice misses most of what the model asks for. FreeToken measures both bandwidths on your machine and splits each step's misses between the two paths in proportion. the GPU and CPU results then merge exactly, with no approximation. two machines with the same GPU can end up wanting opposite strategies, which I did not expect. a 5090 in a gaming desktop should push nearly everything over PCIe, while an 8GB laptop is better off computing most misses on the CPU. none of that is readable off a spec sheet, so the engine profiles it once per machine. the second half of the design is about agents. coding agents constantly rewrite their own history, and every edit normally forces thousands of tokens back through prefill. FreeToken saves its checkpoints at the exact boundaries agent frameworks cut on, so it only reprocesses the new part. its slowest first token stays under 44 seconds, while llama.cpp peaks at 232 and KTransformers at 946. it serves the OpenAI and Anthropic APIs under Apache 2.0, so Claude Code and Codex can point at it directly. releasing weights publicly decides who can download a model, not who can afford to run one. frontier open models keep shipping, and running them still assumes a rented cluster. meanwhile there are over a hundred million consumer machines with discrete GPUs sitting mostly idle. closing that gap was never a hardware problem, and work like this is what turns open weights into something you can actually use. paper: repo: almost every idea in this post, from why memory bandwidth decides the outcome to why moving weights costs more than computing on them, comes straight out of how a GPU is built. I wrote a detailed primer on that. the article is quoted below.
0:42

UC Berkeley just open-sourced FreeToken. (2–4x faster local LLM inference than Ollama) the results are wild: - Qwen3.6-35B on an 8GB GPU at 39.3 tokens/s - DeepSeek-V4-Flash 284B on a 32GB GPU at 22 tokens/s - GLM-5.2 753B on a 96GB GPU at 14.9 tokens/s a 35B model at 16-bit precision needs about 70GB just for its weights. even at 4 bits it is close to 18GB, and FreeToken serves it on an 8GB GPU. let me explain how: all three models mentioned above are Mixture-of-Experts, and that is what FreeToken takes advantage of. each layer holds hundreds of separate experts plus a small router that picks a few of them per token. Qwen3.6-35B activates roughly 3B of its 35B parameters per token. DeepSeek-V4-Flash picks 6 of 256 experts per layer, so 13B of its 284B run at a time. so compute was never the bottleneck. the weights a single step touches fit comfortably on a consumer GPU. every expert the router might pick still has to exist somewhere. they sit in system RAM, and the GPU keeps a cache of the ones the model has been using recently. so everything comes down to what happens when the router picks an expert that is not on the GPU. there are two ways to serve that miss: 1. copy it over PCIe and run it on the GPU 2. run it on the CPU, where it already lives both read from the same system memory, so they compete for one pool of bandwidth instead of adding to each other. existing engines pick one option and freeze it when the model loads. but routing changes on every token, so a fixed choice misses most of what the model asks for. FreeToken measures both bandwidths on your machine and splits each step's misses between the two paths in proportion. the GPU and CPU results then merge exactly, with no approximation. two machines with the same GPU can end up wanting opposite strategies, which I did not expect. a 5090 in a gaming desktop should push nearly everything over PCIe, while an 8GB laptop is better off computing most misses on the CPU. none of that is readable off a spec sheet, so the engine profiles it once per machine. the second half of the design is about agents. coding agents constantly rewrite their own history, and every edit normally forces thousands of tokens back through prefill. FreeToken saves its checkpoints at the exact boundaries agent frameworks cut on, so it only reprocesses the new part. its slowest first token stays under 44 seconds, while llama.cpp peaks at 232 and KTransformers at 946. it serves the OpenAI and Anthropic APIs under Apache 2.0, so Claude Code and Codex can point at it directly. releasing weights publicly decides who can download a model, not who can afford to run one. frontier open models keep shipping, and running them still assumes a rented cluster. meanwhile there are over a hundred million consumer machines with discrete GPUs sitting mostly idle. closing that gap was never a hardware problem, and work like this is what turns open weights into something you can actually use. paper: repo: almost every idea in this post, from why memory bandwidth decides the outcome to why moving weights costs more than computing on them, comes straight out of how a GPU is built. I wrote a detailed primer on that. the article is quoted below.

Akshay 🚀

338,326 Aufrufe • vor 13 Tagen

This is why Nebius will be a trillion dollar hyperscaler (Save this). Nebius is not building another GPU rental shop but rather building a vertically integrated hyperscaler that owns everything from the physical data center, to the server rack hardware it designs in house, to the software stack, to the inference delivery layer. Nearly every other neocloud is essentially a reseller of someone else's infrastructure but Nebius owns the full stack end to end and that distinction is the entire thesis. Here is why vertical integration is the winning architecture for the inference era. AWS and Azure were architected for general purpose computing and every AI workload they run sits on top of infrastructure that was never designed for it, patched, adapted and optimized after the fact. Nebius was built from day one specifically for AI which means every layer of the stack is purpose built and co optimized. The rack design, the networking topology, the cooling systems and the software that orchestrates it all are engineered together as a single system rather than assembled from parts that were never meant to work together. That architectural difference compounds with every passing quarter as AI workloads grow more complex and the performance gap between purpose built and general purpose infrastructure widens. The software layer is where the real competitive moat lives. Most infrastructure companies think of software as a wrapper around hardware while Nebius thinks of software as the product with hardware as the substrate it controls. The company is building an AI native cloud platform where the software layer handles model serving, inference optimization, fine tuning pipelines and developer tooling as first-class primitives. This matters because inference efficiency is almost entirely a software problem. Two companies running identical GPUs can deliver dramatically different performance and cost per token depending on how intelligently the software schedules, batches and routes inference requests across the cluster. Nebius is also building for a fundamental shift in how AI infrastructure gets consumed. Today, enterprise developers navigate massive cloud service catalogs spinning up clusters, managing configurations and building deep expertise in AWS or GCP-specific tooling. The next generation of builders will simply provision agents to interface with infrastructure directly. Nebius is architecting its software layer for that future , one where the interface between the developer and the compute abstraction layer looks nothing like what AWS built in 2006. The entire available capacity has been sold out every quarter. And that is the best possible validation that what Nebius is building is exactly what the market needs and that the market is willing to commit at a scale that makes the current valuation look like the beginning of a much longer story. Long Nebius and make sure to follow me Melvin for more overlooked AI stocks.
4:44

This is why Nebius will be a trillion dollar hyperscaler (Save this). Nebius is not building another GPU rental shop but rather building a vertically integrated hyperscaler that owns everything from the physical data center, to the server rack hardware it designs in house, to the software stack, to the inference delivery layer. Nearly every other neocloud is essentially a reseller of someone else's infrastructure but Nebius owns the full stack end to end and that distinction is the entire thesis. Here is why vertical integration is the winning architecture for the inference era. AWS and Azure were architected for general purpose computing and every AI workload they run sits on top of infrastructure that was never designed for it, patched, adapted and optimized after the fact. Nebius was built from day one specifically for AI which means every layer of the stack is purpose built and co optimized. The rack design, the networking topology, the cooling systems and the software that orchestrates it all are engineered together as a single system rather than assembled from parts that were never meant to work together. That architectural difference compounds with every passing quarter as AI workloads grow more complex and the performance gap between purpose built and general purpose infrastructure widens. The software layer is where the real competitive moat lives. Most infrastructure companies think of software as a wrapper around hardware while Nebius thinks of software as the product with hardware as the substrate it controls. The company is building an AI native cloud platform where the software layer handles model serving, inference optimization, fine tuning pipelines and developer tooling as first-class primitives. This matters because inference efficiency is almost entirely a software problem. Two companies running identical GPUs can deliver dramatically different performance and cost per token depending on how intelligently the software schedules, batches and routes inference requests across the cluster. Nebius is also building for a fundamental shift in how AI infrastructure gets consumed. Today, enterprise developers navigate massive cloud service catalogs spinning up clusters, managing configurations and building deep expertise in AWS or GCP-specific tooling. The next generation of builders will simply provision agents to interface with infrastructure directly. Nebius is architecting its software layer for that future , one where the interface between the developer and the compute abstraction layer looks nothing like what AWS built in 2006. The entire available capacity has been sold out every quarter. And that is the best possible validation that what Nebius is building is exactly what the market needs and that the market is willing to commit at a scale that makes the current valuation look like the beginning of a much longer story. Long Nebius and make sure to follow me Melvin for more overlooked AI stocks.

Melvin

34,306 Aufrufe • vor 1 Monat

QuickSwap is now live on X Layer Mainnet ⚡️ The birth of a new Citadel on the dragon DEX, expanding the Polygon CDK family. Swap and LP on X Layer (farms coming soon) via QuickSwap to experience DeFi on the network at its very best. X Layer is ZK-powered Layer 2 network designed to connect the OKX and Ethereum communities, built using the Polygon CDK tech stack. It's designed to have seamless interoperability with other Layer 2 chains. Now plugged into the AggLayer, OKX’s massive community of 50+ million users is fully interoperable with all interconnected chains and can now tap into its unified liquidity. The network also uses zk-SNARK to make transactions significantly cheaper and far more compact. X Layer brings other benefits to users, such as near-instant finality, unified liquidity, robust security, high scalability... the list goes on. With this mainnet deployment and more to come, QuickSwap is leading the charge to ignite activity and growth on X Layer at rapid speeds. Going full throttle to further unite the Polygon | POL and OKX communities. This is only the beginning of a bright future for QuickSwap, X Layer, and the broader Polygon and Ethereum ecosystems. Get ready to rumble.
0:27

QuickSwap is now live on X Layer Mainnet ⚡️ The birth of a new Citadel on the dragon DEX, expanding the Polygon CDK family. Swap and LP on X Layer (farms coming soon) via QuickSwap to experience DeFi on the network at its very best. X Layer is ZK-powered Layer 2 network designed to connect the OKX and Ethereum communities, built using the Polygon CDK tech stack. It's designed to have seamless interoperability with other Layer 2 chains. Now plugged into the AggLayer, OKX’s massive community of 50+ million users is fully interoperable with all interconnected chains and can now tap into its unified liquidity. The network also uses zk-SNARK to make transactions significantly cheaper and far more compact. X Layer brings other benefits to users, such as near-instant finality, unified liquidity, robust security, high scalability... the list goes on. With this mainnet deployment and more to come, QuickSwap is leading the charge to ignite activity and growth on X Layer at rapid speeds. Going full throttle to further unite the Polygon | POL and OKX communities. This is only the beginning of a bright future for QuickSwap, X Layer, and the broader Polygon and Ethereum ecosystems. Get ready to rumble.

QuickSwap 🐲 DragonFi 2.0

20,857 Aufrufe • vor 2 Jahren