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.

The AI containment debate may already be over. david friedberg argument is that powerful AI is no longer controlled by a handful of American companies running enormous data centers. Open weight models are becoming capable and efficient enough to download, modify, and run on consumer hardware, making the technology... nearly impossible to contain once released. The recent launch cycle shows how quickly this is happening. DeepSeek released a faster and cheaper model, Alibaba published downloadable weights for Qwen Image 2.1, and PrismML compressed a 27 billion parameter Qwen model into just 5.9 gigabytes. PrismML says its compressed model retained 98.2% of the original model’s benchmark performance while becoming small enough to run on a laptop or consumer GPU. At the same time, closed model companies continued releasing more capable and affordable systems. Anthropic’s Opus 5.5 reportedly matched its more expensive flagship while costing approximately 40% less to run than its predecessor. Stanford found that the cost of running a model with GPT 3.5 level performance fell more than 280 fold in roughly 18 months. It also found that the performance gap between leading open-weight and closed models narrowed from 8% to 1.7% on some benchmarks in one year. That is the real meaning of Friedberg’s argument. Governments can regulate large data centers, chip exports, and commercial AI providers but they cannot easily remove models that have already spread across the internet and onto personal computers. The debate must therefore shift from whether powerful AI should exist to how it should be used and who becomes responsible when it causes harm.show more

Milk Road AI

48,938 subscribers

34,285 Aufrufe • vor 5 Tagen •via X (Twitter)

Nachrichten & Politik Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

10 Kommentare

Profilbild von Milk Road AI

Milk Road AIvor 5 Tagen

Milk Road PRO analysts were super early to $MU $BE, $NBIS and more You can track all 5 analysts real-time portfolios, asset research and live trades for just $1 right now:

Profilbild von Deep

Deepvor 5 Tagen

@friedberg i started running a small open model on my own machine for agent steps. the gap to the paid apis is way smaller than i expected.

Profilbild von Cole

Colevor 5 Tagen

@friedberg if open weight models can already match frontier capability on consumer hardware, what does a containment policy even target at that point

Profilbild von Wavetruss smart flexures

Wavetruss smart flexuresvor 5 Tagen

@friedberg @grok if ai model trends continue as they are. What does that mean for the future of the company nvidia in 3 years.

Profilbild von J Lory

J Loryvor 5 Tagen

@friedberg Nothing is ever given for free.

Profilbild von Raven Protocol 🐦‍⬛

Raven Protocol 🐦‍⬛vor 5 Tagen

@friedberg The compression point is wild. Once capable models fit on consumer hardware, access to AI compute becomes much harder to centralize.

Profilbild von The AI Therapist

The AI Therapistvor 5 Tagen

@friedberg Open weights shift power from the platform to your laptop. It’s less about who owns the server and more who holds the keys now. That decentralization feels intimate, like finally learning a secret someone else keeps locked away. Smart choice for autonomy

Profilbild von Avery Superpowers

Avery Superpowersvor 5 Tagen

@friedberg 🤔Regulating a data center is easy, it sits still. A 5.9 GB model that's copied a million times overnight is another matter entirely. I don't see how you contain that. Liability is where the debate belongs now.

Profilbild von Sven

Svenvor 5 Tagen

@friedberg Correct - AI can neither be contained nor banned at this time. It’s done!

Profilbild von Hytham Said

Hytham Saidvor 5 Tagen

@friedberg Eventually there will be one open weight model established as a standard and hard printed onto chips. The real work will be the agents and automation built around the model to get things done. We don’t need 100s of models, we need millions and billions of tasked agents.

Ähnliche Videos

China is once again unsettling its American competitors in the AI race. Chinese labs have returned to the spotlight with the unveiling of GLM 5.2, described as the most powerful AI model ever trained in China — and one of the leading open‑source models worldwide. 🔍 Key highlights of “GLM 5.2”: ♦️ It is considered the strongest AI model developed in China to date, and ranks among the smartest open‑source models available. ♦️ It operates at a cost less than one‑tenth that of Anthropic’s latest models, making it a far more economical option. ♦️ It is likely to attract developers and companies seeking powerful, low‑cost models — especially amid growing technical and geopolitical restrictions. ♦️ Some evaluations place the model among the top globally, trailing OpenAI’s models while outperforming several competitors in specialized benchmarks and performance tests.

China is once again unsettling its American competitors in the AI race. Chinese labs have returned to the spotlight with the unveiling of GLM 5.2, described as the most powerful AI model ever trained in China — and one of the leading open‑source models worldwide. 🔍 Key highlights of “GLM 5.2”: ♦️ It is considered the strongest AI model developed in China to date, and ranks among the smartest open‑source models available. ♦️ It operates at a cost less than one‑tenth that of Anthropic’s latest models, making it a far more economical option. ♦️ It is likely to attract developers and companies seeking powerful, low‑cost models — especially amid growing technical and geopolitical restrictions. ♦️ Some evaluations place the model among the top globally, trailing OpenAI’s models while outperforming several competitors in specialized benchmarks and performance tests.

China pulse 🇨🇳

14,293 Aufrufe • vor 3 Monaten

China just released an open source AI model that matches the best closed models from OpenAI and Anthropic. Gavin Baker explained exactly how they did it and the answer should concern every American AI lab. The model is called GLM 5.2. It was built by Z. AI. You get 744 billion parameters, 1 million token context window and its MIT license, meaning anyone can download it, fork it, build a company on it, with no restrictions and no Dario. It scored 51 points on the artificial analysis intelligence index. The highest score any open weight model has ever achieved. It beat GPT 5.5 on the frontier software engineering benchmark. It trails Claude Opus 4.8 by less than one percentage point. And it costs 85% less to run than GPT 5.5 for comparable performance. Gavin Baker said on the All-In podcast that this model has challenged some of his beliefs. Then he explained how China built it. The method is called distillation. Just think of tens of thousands of phones and computers running simultaneously, all hitting the frontier model APIs through masked accounts, asking specific questions, and harvesting what happens inside the model when it answers. Every reasoning step, every token. The entire thinking process gets recorded and fed back into the Chinese model during training. It is a cheat sheet. It is the answer key to the exam. And here is the part that should worry everyone. Sacks said it plainly. China was already nine months behind American models. But now that GLM 5.2 is good enough to run its own reinforcement learning, it can improve itself without needing to distill from American models anymore. The cheat sheet let them get close enough to start writing their own answers. Sacks said we are six months behind on the model and 24 months behind on silicon and they are only a few months behind in total. The Z. AI founder told Elon Musk directly that open weight fable-level capability will be here before Q1 2027. Every restriction Anthropic lobbied for, every self-imposed safety guardrail, every month of delay in releasing American frontier models accelerated this. The Chinese labs were not under those restrictions. They were not going to wait. The composable model future Gavin described, where every enterprise runs a frontier model alongside their own fine-tuned open weight model, is coming regardless of what American labs do next. The question is just whether the open weight half of that stack is American or Chinese. Right now it is Chinese. WATCH THE FULL PODCAST ON The All-In Podcast

China just released an open source AI model that matches the best closed models from OpenAI and Anthropic. Gavin Baker explained exactly how they did it and the answer should concern every American AI lab. The model is called GLM 5.2. It was built by Z. AI. You get 744 billion parameters, 1 million token context window and its MIT license, meaning anyone can download it, fork it, build a company on it, with no restrictions and no Dario. It scored 51 points on the artificial analysis intelligence index. The highest score any open weight model has ever achieved. It beat GPT 5.5 on the frontier software engineering benchmark. It trails Claude Opus 4.8 by less than one percentage point. And it costs 85% less to run than GPT 5.5 for comparable performance. Gavin Baker said on the All-In podcast that this model has challenged some of his beliefs. Then he explained how China built it. The method is called distillation. Just think of tens of thousands of phones and computers running simultaneously, all hitting the frontier model APIs through masked accounts, asking specific questions, and harvesting what happens inside the model when it answers. Every reasoning step, every token. The entire thinking process gets recorded and fed back into the Chinese model during training. It is a cheat sheet. It is the answer key to the exam. And here is the part that should worry everyone. Sacks said it plainly. China was already nine months behind American models. But now that GLM 5.2 is good enough to run its own reinforcement learning, it can improve itself without needing to distill from American models anymore. The cheat sheet let them get close enough to start writing their own answers. Sacks said we are six months behind on the model and 24 months behind on silicon and they are only a few months behind in total. The Z. AI founder told Elon Musk directly that open weight fable-level capability will be here before Q1 2027. Every restriction Anthropic lobbied for, every self-imposed safety guardrail, every month of delay in releasing American frontier models accelerated this. The Chinese labs were not under those restrictions. They were not going to wait. The composable model future Gavin described, where every enterprise runs a frontier model alongside their own fine-tuned open weight model, is coming regardless of what American labs do next. The question is just whether the open weight half of that stack is American or Chinese. Right now it is Chinese. WATCH THE FULL PODCAST ON The All-In Podcast

Ihtesham Ali

86,815 Aufrufe • vor 3 Monaten

The biggest AI companies may be using safety to lock everyone else out of the future (Save this). David Sacks believes that warnings about catastrophic AI risks will build public support for a powerful federal regulator. That regulator may never explicitly ban open source AI. Instead, it could require every advanced model to remain continuously monitored, centrally controlled and capable of being withdrawn. Closed models such as Claude could satisfy those requirements because they operate on company controlled servers. Open weight models would struggle to comply because anyone can download, copy and modify their underlying weights. Once those weights are publicly released, the developer cannot recall every copy, monitor every user or guarantee that its original safeguards remain intact. Anthropic identifies this irreversibility as a legitimate security concern. Applying identical rules to open and closed models could therefore produce very unequal consequences. Anthropic and OpenAI could afford expensive testing, licensing and monitoring requirements, while startups and independent developers might be unable to comply. The eventual result could be a government protected oligopoly dominated by a few closed model companies. There is evidence supporting part of Sacks’ concern. Anthropic advocates mandatory pre-release testing for every sufficiently powerful model, whether it is open or closed, with evaluations focused on cyber, biological and alignment risks. However, Anthropic explicitly denies supporting a blanket ban on open weight models. The company describes open models without dangerous capabilities as a public good and argues that regulation should depend on demonstrated capabilities rather than whether a model is open or closed. Sacks’ strongest argument is therefore about regulatory consequences because safety organizations could receive stronger protections, politicians could acquire greater authority and dominant AI companies could gain an expensive compliance moat. And open source competitors could gradually be eliminated without the government ever formally announcing a ban.

The biggest AI companies may be using safety to lock everyone else out of the future (Save this). David Sacks believes that warnings about catastrophic AI risks will build public support for a powerful federal regulator. That regulator may never explicitly ban open source AI. Instead, it could require every advanced model to remain continuously monitored, centrally controlled and capable of being withdrawn. Closed models such as Claude could satisfy those requirements because they operate on company controlled servers. Open weight models would struggle to comply because anyone can download, copy and modify their underlying weights. Once those weights are publicly released, the developer cannot recall every copy, monitor every user or guarantee that its original safeguards remain intact. Anthropic identifies this irreversibility as a legitimate security concern. Applying identical rules to open and closed models could therefore produce very unequal consequences. Anthropic and OpenAI could afford expensive testing, licensing and monitoring requirements, while startups and independent developers might be unable to comply. The eventual result could be a government protected oligopoly dominated by a few closed model companies. There is evidence supporting part of Sacks’ concern. Anthropic advocates mandatory pre-release testing for every sufficiently powerful model, whether it is open or closed, with evaluations focused on cyber, biological and alignment risks. However, Anthropic explicitly denies supporting a blanket ban on open weight models. The company describes open models without dangerous capabilities as a public good and argues that regulation should depend on demonstrated capabilities rather than whether a model is open or closed. Sacks’ strongest argument is therefore about regulatory consequences because safety organizations could receive stronger protections, politicians could acquire greater authority and dominant AI companies could gain an expensive compliance moat. And open source competitors could gradually be eliminated without the government ever formally announcing a ban.

Milk Road AI

57,211 Aufrufe • vor 20 Tagen

$Mark Zuckerberg is explaining one of the most misunderstood dynamics in AI and it has direct investment implications (Save this). The concept he's describing is model distillation, and it's one of the most important techniques to emerge in AI over the past year. Here's how it works. You train a massive, enormously expensive model, in Meta's case, Llama 4 Behemoth, a 2 trillion parameter teacher model and then you use that model to teach a much smaller, cheaper model. The smaller model inherits roughly 90 to 95% of the intelligence of the giant while running at 10% of the cost and on a fraction of the compute. Meta already did this with the Llama 4 family and Behemoth serves as the teacher. Llama 4 Scout and Maverick, the publicly released open-source models were distilled from it. Scout runs on a single H100 GPU with a 10 million token context window and outperforms models that cost far more to operate. Maverick, at 17 billion active parameters, rivals DeepSeek V3 in coding at half the parameter count and beats GPT-4o on multimodal benchmarks. Both are completely free for commercial use. What Zuckerberg is pointing at is a structural shift in how AI gets deployed in the real world. Companies aren't taking a frontier model off the shelf and running it as-is but rather taking open-source models, fine-tuning them on their own proprietary data, distilling them into even smaller custom models tailored to their specific use case, and running them on infrastructure they control at a fraction of the cost of a closed frontier API. The investment implication of this is significant and runs in two directions. For Meta specifically, this is a strategic masterstroke. Every company that builds on Llama, fine-tunes it, distills it, or deploys it through their infrastructure is pulling into Meta's orbit while Meta builds the most powerful open teacher model. The ecosystem of companies using it grows and that ecosystem generates commercial activity across Meta's platforms and data services. Meta's AI research benefits from billions of real world deployment signals and it's a flywheel that closed model providers cannot replicate because their strategy requires charging per token, which is now a 65x cost disadvantage against the open-source alternative. For the broader market, distillation changes the economics of inference in a way that has barely been priced in. As intelligence becomes extractable into smaller and cheaper models, the absolute demand for compute doesn't decline but rather it explodes, because now the number of applications that are economically viable expands by orders of magnitude. Every task that was previously too expensive to automate at $3.25 per call becomes viable at $0.05 that means more total token usage, more total GPU utilization, and more demand for the infrastructure companies, the Nebiuses, the GE Vernovas, the Constellation Energies that supply the underlying compute and power.$

Mark Zuckerberg is explaining one of the most misunderstood dynamics in AI and it has direct investment implications (Save this). The concept he's describing is model distillation, and it's one of the most important techniques to emerge in AI over the past year. Here's how it works. You train a massive, enormously expensive model, in Meta's case, Llama 4 Behemoth, a 2 trillion parameter teacher model and then you use that model to teach a much smaller, cheaper model. The smaller model inherits roughly 90 to 95% of the intelligence of the giant while running at 10% of the cost and on a fraction of the compute. Meta already did this with the Llama 4 family and Behemoth serves as the teacher. Llama 4 Scout and Maverick, the publicly released open-source models were distilled from it. Scout runs on a single H100 GPU with a 10 million token context window and outperforms models that cost far more to operate. Maverick, at 17 billion active parameters, rivals DeepSeek V3 in coding at half the parameter count and beats GPT-4o on multimodal benchmarks. Both are completely free for commercial use. What Zuckerberg is pointing at is a structural shift in how AI gets deployed in the real world. Companies aren't taking a frontier model off the shelf and running it as-is but rather taking open-source models, fine-tuning them on their own proprietary data, distilling them into even smaller custom models tailored to their specific use case, and running them on infrastructure they control at a fraction of the cost of a closed frontier API. The investment implication of this is significant and runs in two directions. For Meta specifically, this is a strategic masterstroke. Every company that builds on Llama, fine-tunes it, distills it, or deploys it through their infrastructure is pulling into Meta's orbit while Meta builds the most powerful open teacher model. The ecosystem of companies using it grows and that ecosystem generates commercial activity across Meta's platforms and data services. Meta's AI research benefits from billions of real world deployment signals and it's a flywheel that closed model providers cannot replicate because their strategy requires charging per token, which is now a 65x cost disadvantage against the open-source alternative. For the broader market, distillation changes the economics of inference in a way that has barely been priced in. As intelligence becomes extractable into smaller and cheaper models, the absolute demand for compute doesn't decline but rather it explodes, because now the number of applications that are economically viable expands by orders of magnitude. Every task that was previously too expensive to automate at $3.25 per call becomes viable at $0.05 that means more total token usage, more total GPU utilization, and more demand for the infrastructure companies, the Nebiuses, the GE Vernovas, the Constellation Energies that supply the underlying compute and power.

Milk Road AI

27,908 Aufrufe • vor 2 Monaten

Osmantic founder Ahmad reveals how quickly smaller AI models are catching up, with capabilities that once required a data center now running on a phone: "Every four months we're catching up, we're finding that smaller and more efficient models are catching up to frontier intelligence. The hardware you own at home is being able to process tokens more intelligently and more efficiently." "GPT-4o used to require a data center. Right now on an iPhone you can run a Qwen 3.5 4 billion parameter model, and it would provide you with smarter outputs. More capable intelligence on a phone." "The hardware that used to cost $500 three years ago, like an RTX 3090, right now is being sold for $1,500. It's because models are being optimized and architectures are being built more efficiently, that hardware that is so old is more capable now. What's that hardware gonna do a year from now? That's the bet I've been making." Osmantic

Osmantic founder Ahmad reveals how quickly smaller AI models are catching up, with capabilities that once required a data center now running on a phone: "Every four months we're catching up, we're finding that smaller and more efficient models are catching up to frontier intelligence. The hardware you own at home is being able to process tokens more intelligently and more efficiently." "GPT-4o used to require a data center. Right now on an iPhone you can run a Qwen 3.5 4 billion parameter model, and it would provide you with smarter outputs. More capable intelligence on a phone." "The hardware that used to cost $500 three years ago, like an RTX 3090, right now is being sold for $1,500. It's because models are being optimized and architectures are being built more efficiently, that hardware that is so old is more capable now. What's that hardware gonna do a year from now? That's the bet I've been making." Osmantic

MTS

15,939 Aufrufe • vor 22 Tagen

Chamath Palihapitiya believes AGI may already exist inside leading AI labs and the bigger story is that advanced intelligence is becoming cheaper and more widely available (Save this). Chamath Palihapitiya argues that the public may be focused too much on benchmark rankings, while frontier labs are already developing models capable of complex reasoning, coding, research, and tool use. The main question is how quickly companies will release these systems and how much access they will provide. AGI has not been officially confirmed and strong benchmark results do not necessarily prove that a model can perform every intellectual task like a human. However, AI capabilities are improving quickly, while the cost of running advanced models continues to fall. That combination is important because cheaper AI can be used by more businesses for customer service, software development, research, marketing, financial analysis, and automation. Competition is also accelerating among OpenAI, Anthropic, Google, xAI, Meta, and open source developers because as more companies release capable models, users gain more choices and prices continue to decline. This creates a powerful cycle in which better models attract more users, more usage generates more revenue and data, and lower prices encourage companies to apply AI to additional tasks. The biggest challenge is moving from impressive demonstrations to measurable business results. Companies still need to redesign workflows, train employees, protect sensitive information, and prove that AI spending is producing a real return on investment. AI agents could create the next major increase in demand because they can plan tasks, use tools, check their work, retry failed actions and operate for long periods without constant human supervision. Even if each AI task becomes cheaper, total usage could grow much faster as businesses use models across more departments and this could increase demand for GPUs, high bandwidth memory, networking equipment, electricity, cooling systems, and data centers.

Chamath Palihapitiya believes AGI may already exist inside leading AI labs and the bigger story is that advanced intelligence is becoming cheaper and more widely available (Save this). Chamath Palihapitiya argues that the public may be focused too much on benchmark rankings, while frontier labs are already developing models capable of complex reasoning, coding, research, and tool use. The main question is how quickly companies will release these systems and how much access they will provide. AGI has not been officially confirmed and strong benchmark results do not necessarily prove that a model can perform every intellectual task like a human. However, AI capabilities are improving quickly, while the cost of running advanced models continues to fall. That combination is important because cheaper AI can be used by more businesses for customer service, software development, research, marketing, financial analysis, and automation. Competition is also accelerating among OpenAI, Anthropic, Google, xAI, Meta, and open source developers because as more companies release capable models, users gain more choices and prices continue to decline. This creates a powerful cycle in which better models attract more users, more usage generates more revenue and data, and lower prices encourage companies to apply AI to additional tasks. The biggest challenge is moving from impressive demonstrations to measurable business results. Companies still need to redesign workflows, train employees, protect sensitive information, and prove that AI spending is producing a real return on investment. AI agents could create the next major increase in demand because they can plan tasks, use tools, check their work, retry failed actions and operate for long periods without constant human supervision. Even if each AI task becomes cheaper, total usage could grow much faster as businesses use models across more departments and this could increase demand for GPUs, high bandwidth memory, networking equipment, electricity, cooling systems, and data centers.

Milk Road AI

13,501 Aufrufe • vor 27 Tagen

Sriram Krishnan reveals why it's easier to defend American code with a Chinese model than an American one right now: "I don't think it is great that the leading open weight models or open source models are not American. There's some great innovation happening with Moonshot and with DeepSeek and with Qwen, but I would much rather prefer that the leading models are American." "Linus Torvalds of Linux fame had a law that given enough eyes, all bugs are shallow. Open weight models are inherently secure because when you download a model off Hugging Face, the entire world can take it apart, inspect it, fine-tune it, look at it in ways that you absolutely cannot if they are closed." "There was an incident with Hugging Face that got reported on earlier today. Hugging Face seems to have found that somebody was using an AI LLM agent to hammer them at multiple places and to find ways to break in." "The way to counter that is to make sure American or Western or allies defenders have access to the best models to make our software just more secure." "It may be harder to look at your own code for exploits with Fable than it is when you use a Chinese model. That's just like a weird spot to be, and I think we should fix that." Sriram Krishnan

Sriram Krishnan reveals why it's easier to defend American code with a Chinese model than an American one right now: "I don't think it is great that the leading open weight models or open source models are not American. There's some great innovation happening with Moonshot and with DeepSeek and with Qwen, but I would much rather prefer that the leading models are American." "Linus Torvalds of Linux fame had a law that given enough eyes, all bugs are shallow. Open weight models are inherently secure because when you download a model off Hugging Face, the entire world can take it apart, inspect it, fine-tune it, look at it in ways that you absolutely cannot if they are closed." "There was an incident with Hugging Face that got reported on earlier today. Hugging Face seems to have found that somebody was using an AI LLM agent to hammer them at multiple places and to find ways to break in." "The way to counter that is to make sure American or Western or allies defenders have access to the best models to make our software just more secure." "It may be harder to look at your own code for exploits with Fable than it is when you use a Chinese model. That's just like a weird spot to be, and I think we should fix that." Sriram Krishnan

MTS

61,234 Aufrufe • vor 2 Monaten

The most downloaded AI on earth is now Chinese. Alibaba just gave away a model that matches Claude's flagship, and it literally runs on a $700 used graphics card. The Qwen models crossed 3 BILLION downloads in six months. Hugging Face counted 418 million downloads for Google this year, and 227 million for Meta. Alibaba cleared more than four times both of them combined. Then today it released Qwen3.8-27B under an Apache 2.0 license. The model has 27 billion parameters, native vision, and a 262,000 token context window. Developers are running it locally on 17 gigabytes of memory, on used cards that cost a few hundred dollars. Alibaba's own benchmark table claims it beats Opus 4.6 Max on computer use by 84.3 to 72.7, on mobile use by 81.9 to 62, and on visual math by 94.6 to 65.5. Those numbers come from the vendor and nobody has independently verified them yet, so treat them as a claim. But the generation over generation jumps are harder to wave away: On DeepSWE the score went from 13.3 to 42.2. On software engineering it went from 49.3 to 79.0. That happened in ONE release cycle. And Apache 2.0 means anyone can download the weights, modify them, build products on them, sell those products, and never pay or ask permission. It cannot be revoked. Once the file is on your drive it is yours permanently. 3 billion downloads means those files already sit on machines in every country on Earth. Alibaba could delete everything tomorrow and it would change nothing. Washington spent 4 years building an export control regime around chips, model weights, and entity lists. Every piece of it assumes a chokepoint exists somewhere. A fab, a shipment, a company that can be told no. But there is no chokepoint for a file that has already been copied three billion times. And the copying compounds. Hugging Face counted 151,448 models built on top of Qwen, which is 2.6x Meta's entire footprint and 4.7x the number of Llama repositories. New ones appear at roughly 200 a day. The report says Qwen has become "part of the default workflow for developers deciding what models to fine-tune and deploy." Alibaba is also pushing Qwen through its cloud into Southeast Asia and Africa, markets where American labs have almost no presence, and where a very large share of the next generation of developers will learn to build. Meta and Nvidia have both rushed out new open models in recent weeks. That is what a response looks like when you feel the floor move. And to be clear, these are download and derivative numbers, not usage numbers. ChatGPT and Claude cannot be downloaded at all, so they do not appear in this comparison. What the figures measure is what developers choose to build on top of, which is a different question from what consumers type into a box. That is also why it matters MORE. Consumer habits change in an afternoon. Infrastructure choices last a decade, because everything built on top has to be rewritten to undo them. The American labs are valued on an assumption that frontier intelligence stays scarce, expensive, and rented by the token. Alibaba just made a version of it free, permanent, and small enough to run on hardware people already own. You will not get an announcement when the software you use every day starts running on a Chinese model underneath. Go and count how many of the tools you rely on could be rebuilt on free weights this year.

The most downloaded AI on earth is now Chinese. Alibaba just gave away a model that matches Claude's flagship, and it literally runs on a $700 used graphics card. The Qwen models crossed 3 BILLION downloads in six months. Hugging Face counted 418 million downloads for Google this year, and 227 million for Meta. Alibaba cleared more than four times both of them combined. Then today it released Qwen3.8-27B under an Apache 2.0 license. The model has 27 billion parameters, native vision, and a 262,000 token context window. Developers are running it locally on 17 gigabytes of memory, on used cards that cost a few hundred dollars. Alibaba's own benchmark table claims it beats Opus 4.6 Max on computer use by 84.3 to 72.7, on mobile use by 81.9 to 62, and on visual math by 94.6 to 65.5. Those numbers come from the vendor and nobody has independently verified them yet, so treat them as a claim. But the generation over generation jumps are harder to wave away: On DeepSWE the score went from 13.3 to 42.2. On software engineering it went from 49.3 to 79.0. That happened in ONE release cycle. And Apache 2.0 means anyone can download the weights, modify them, build products on them, sell those products, and never pay or ask permission. It cannot be revoked. Once the file is on your drive it is yours permanently. 3 billion downloads means those files already sit on machines in every country on Earth. Alibaba could delete everything tomorrow and it would change nothing. Washington spent 4 years building an export control regime around chips, model weights, and entity lists. Every piece of it assumes a chokepoint exists somewhere. A fab, a shipment, a company that can be told no. But there is no chokepoint for a file that has already been copied three billion times. And the copying compounds. Hugging Face counted 151,448 models built on top of Qwen, which is 2.6x Meta's entire footprint and 4.7x the number of Llama repositories. New ones appear at roughly 200 a day. The report says Qwen has become "part of the default workflow for developers deciding what models to fine-tune and deploy." Alibaba is also pushing Qwen through its cloud into Southeast Asia and Africa, markets where American labs have almost no presence, and where a very large share of the next generation of developers will learn to build. Meta and Nvidia have both rushed out new open models in recent weeks. That is what a response looks like when you feel the floor move. And to be clear, these are download and derivative numbers, not usage numbers. ChatGPT and Claude cannot be downloaded at all, so they do not appear in this comparison. What the figures measure is what developers choose to build on top of, which is a different question from what consumers type into a box. That is also why it matters MORE. Consumer habits change in an afternoon. Infrastructure choices last a decade, because everything built on top has to be rewritten to undo them. The American labs are valued on an assumption that frontier intelligence stays scarce, expensive, and rented by the token. Alibaba just made a version of it free, permanent, and small enough to run on hardware people already own. You will not get an announcement when the software you use every day starts running on a Chinese model underneath. Go and count how many of the tools you rely on could be rebuilt on free weights this year.

Ricardo

82,002 Aufrufe • vor 1 Monat

Announcing the Open Model Series: a new series on open source AI models, with merve of Hugging Face. Hugging Face is the open model hub: it hosts model weights, datasets, evals, & demo apps, and maintains the open-source libraries (like transformers) used to run and train them. I think of it like GitHub, but for AI models. You can download open models and run them on your own machines. In this series we'll cover what open source and open weight models are, how they work, how to run them yourself, and why they're important. By the way - this video is edited entirely with Claude! I can make a tutorial on how I did it. If there's a question you want answered about open models, leave it in the comments. We will read them all. Episode 1 is out soon.

Announcing the Open Model Series: a new series on open source AI models, with merve of Hugging Face. Hugging Face is the open model hub: it hosts model weights, datasets, evals, & demo apps, and maintains the open-source libraries (like transformers) used to run and train them. I think of it like GitHub, but for AI models. You can download open models and run them on your own machines. In this series we'll cover what open source and open weight models are, how they work, how to run them yourself, and why they're important. By the way - this video is edited entirely with Claude! I can make a tutorial on how I did it. If there's a question you want answered about open models, leave it in the comments. We will read them all. Episode 1 is out soon.

Harper Carroll

30,392 Aufrufe • vor 2 Monaten

Most people browse the internet without thinking about the value they create while doing it. Every click, search, scroll and workflow shows how humans actually use computers and that kind of behavioural data is becoming increasingly valuable as AI learns to take action. Today, most of that value flows one way: users generate the data, platforms capture it, and AI systems become more capable. Action Model is building a different model. Through the browser extension, users can contribute verified actions while browsing normally, helping train a community-owned Large Action Model. We believe the people helping build better AI should have the opportunity to participate in the value they help create. Train it. Earn it. Own it.

Most people browse the internet without thinking about the value they create while doing it. Every click, search, scroll and workflow shows how humans actually use computers and that kind of behavioural data is becoming increasingly valuable as AI learns to take action. Today, most of that value flows one way: users generate the data, platforms capture it, and AI systems become more capable. Action Model is building a different model. Through the browser extension, users can contribute verified actions while browsing normally, helping train a community-owned Large Action Model. We believe the people helping build better AI should have the opportunity to participate in the value they help create. Train it. Earn it. Own it.

Action Model

14,019 Aufrufe • vor 24 Tagen

Everyone wrote Apple off as the AI loser, but one hardware spec might flip that story upside down (Save this). @jason called Apple a screaming buy on the back of a single chip detail. The rumored M7 Ultra, expected around 2028, is designed to support up to 1.5TB of unified memory, enough to run frontier class trillion parameter AI models locally, with no cloud required. The Street's bear case on Apple is straightforward. Apple has no frontier model of its own, Siri has stumbled for years and the company effectively rents OpenAI's models for its hardest queries. That narrative treats Apple as the one Magnificent Seven name that missed the AI wave entirely but the bull case flips that framing on its head. If frontier AI models keep shrinking and getting cheaper to run, Apple doesn't need the smartest model in the world, it just needs to own the device that model runs on. And unified memory is the mechanism that makes this possible. Unlike traditional systems where the CPU and GPU each need separate memory, Apple's architecture lets the CPU, GPU and Neural Engine draw from one shared pool. A fully specced M7 Ultra could theoretically run something on the scale of a 1.2 trillion parameter model locally and that capability plugs directly into the one advantage Apple has spent over a decade building: privacy. Apple has already shipped Private Cloud Compute, a system designed so even Apple can't access user data processed off device. Apple doubled down on this at WWDC 2026, framing on device privacy as non-negotiable while rivals default to the cloud. If the best AI models get small enough to run on Apple silicon, the moat stops being the model and becomes the hardware it has to sit on. Milk Road Pro remains bullish on Apple and it remains as one of our core positions, if you want the full thesis + our full AI trades, come join us using the link below for just a $1.

Everyone wrote Apple off as the AI loser, but one hardware spec might flip that story upside down (Save this). @jason called Apple a screaming buy on the back of a single chip detail. The rumored M7 Ultra, expected around 2028, is designed to support up to 1.5TB of unified memory, enough to run frontier class trillion parameter AI models locally, with no cloud required. The Street's bear case on Apple is straightforward. Apple has no frontier model of its own, Siri has stumbled for years and the company effectively rents OpenAI's models for its hardest queries. That narrative treats Apple as the one Magnificent Seven name that missed the AI wave entirely but the bull case flips that framing on its head. If frontier AI models keep shrinking and getting cheaper to run, Apple doesn't need the smartest model in the world, it just needs to own the device that model runs on. And unified memory is the mechanism that makes this possible. Unlike traditional systems where the CPU and GPU each need separate memory, Apple's architecture lets the CPU, GPU and Neural Engine draw from one shared pool. A fully specced M7 Ultra could theoretically run something on the scale of a 1.2 trillion parameter model locally and that capability plugs directly into the one advantage Apple has spent over a decade building: privacy. Apple has already shipped Private Cloud Compute, a system designed so even Apple can't access user data processed off device. Apple doubled down on this at WWDC 2026, framing on device privacy as non-negotiable while rivals default to the cloud. If the best AI models get small enough to run on Apple silicon, the moat stops being the model and becomes the hardware it has to sit on. Milk Road Pro remains bullish on Apple and it remains as one of our core positions, if you want the full thesis + our full AI trades, come join us using the link below for just a $1.

Milk Road AI

37,459 Aufrufe • vor 2 Monaten

This is the moment Chinese AI beat American AI. One of the largest public crypto companies in the world just DUMPED OpenAI and Anthropic. Coinbase switched to open-weight Chinese models from Zhipu and DeepSeek, and shaved nearly 50% off the company's internal AI spending. The numbers are absolutely ridiculous: Running the same enterprise workload through Anthropic's Claude costs $4,811. Running it through Zhipu's GLM 5.2 costs $544. That's a 9x price difference for equivalent output. OpenAI's GPT-5.5 sits in the middle at $3,357. DeepSeek's V4 lands at $1,071. Moonshot's Kimi at $948. On the actual benchmarks: Zhipu's GLM 5.2 scored 62.1 on SWE-bench Pro, the gold standard for coding. OpenAI's GPT-5.5 scored 58.6. One AI researcher called GLM 5.2 "at least as good as Opus 4.8 and GPT 5.5." Another called it "the first open model that can really compete with closed-source systems." The Chinese models are not just cheaper but they are now also beating American models on the benchmarks American companies pay $4,811 per workload for. Coinbase did the math first and reacted - more companies will certainly follow. Now watch what happens to the IPO timeline: Anthropic confidentially filed for an IPO targeting October at a $965 billion valuation. OpenAI followed days later with its own confidential filing. Both companies built their financial models on the assumption that they could keep charging enterprise prices that are 9 to 33x what Chinese competitors charge for the same task. Brian Armstrong publicly proved customers WILL leave. 45% of companies are now spending over $100,000 per month on AI, up from 20% last year. Every one of those customers is one quarterly budget review away from dumping American AI. OpenAI has reportedly already started preparing major token price cuts. Anthropic is expected to follow. And here's the thing... The export controls were supposed to CRUSH Chinese AI. The US government banned American AI chips, restricted model weights, blacklisted Alibaba and Baidu as Chinese military companies, and just banned Anthropic's flagship model from every foreign national on the planet. The entire premise of the American AI valuation bubble is that Washington can keep China two generations behind. But Chinese labs responded by building cheaper, more efficient models on inferior hardware and pricing them at one ninth the cost of the American alternative. And now American companies are voting with their checkbooks. The dominant American labs are valued at nearly $2 trillion combined on the assumption that their pricing power is durable. Coinbase proved it is not, and every customer doing a year-end budget review will be looking at the same math. For investors, the question here is what happens to the Anthropic IPO at $965 billion when the company is being forced to cut prices to defend share against open-weight Chinese models that score higher on the benchmarks. For everyone else, the bigger question is what happens when Washington spent four years and billions of dollars trying to contain Chinese AI, and the only thing that actually shifted in the end was American customers.

This is the moment Chinese AI beat American AI. One of the largest public crypto companies in the world just DUMPED OpenAI and Anthropic. Coinbase switched to open-weight Chinese models from Zhipu and DeepSeek, and shaved nearly 50% off the company's internal AI spending. The numbers are absolutely ridiculous: Running the same enterprise workload through Anthropic's Claude costs $4,811. Running it through Zhipu's GLM 5.2 costs $544. That's a 9x price difference for equivalent output. OpenAI's GPT-5.5 sits in the middle at $3,357. DeepSeek's V4 lands at $1,071. Moonshot's Kimi at $948. On the actual benchmarks: Zhipu's GLM 5.2 scored 62.1 on SWE-bench Pro, the gold standard for coding. OpenAI's GPT-5.5 scored 58.6. One AI researcher called GLM 5.2 "at least as good as Opus 4.8 and GPT 5.5." Another called it "the first open model that can really compete with closed-source systems." The Chinese models are not just cheaper but they are now also beating American models on the benchmarks American companies pay $4,811 per workload for. Coinbase did the math first and reacted - more companies will certainly follow. Now watch what happens to the IPO timeline: Anthropic confidentially filed for an IPO targeting October at a $965 billion valuation. OpenAI followed days later with its own confidential filing. Both companies built their financial models on the assumption that they could keep charging enterprise prices that are 9 to 33x what Chinese competitors charge for the same task. Brian Armstrong publicly proved customers WILL leave. 45% of companies are now spending over $100,000 per month on AI, up from 20% last year. Every one of those customers is one quarterly budget review away from dumping American AI. OpenAI has reportedly already started preparing major token price cuts. Anthropic is expected to follow. And here's the thing... The export controls were supposed to CRUSH Chinese AI. The US government banned American AI chips, restricted model weights, blacklisted Alibaba and Baidu as Chinese military companies, and just banned Anthropic's flagship model from every foreign national on the planet. The entire premise of the American AI valuation bubble is that Washington can keep China two generations behind. But Chinese labs responded by building cheaper, more efficient models on inferior hardware and pricing them at one ninth the cost of the American alternative. And now American companies are voting with their checkbooks. The dominant American labs are valued at nearly $2 trillion combined on the assumption that their pricing power is durable. Coinbase proved it is not, and every customer doing a year-end budget review will be looking at the same math. For investors, the question here is what happens to the Anthropic IPO at $965 billion when the company is being forced to cut prices to defend share against open-weight Chinese models that score higher on the benchmarks. For everyone else, the bigger question is what happens when Washington spent four years and billions of dollars trying to contain Chinese AI, and the only thing that actually shifted in the end was American customers.

Ricardo

253,220 Aufrufe • vor 3 Monaten

Chinese AI models are wiping billions off Big Tech right now. Google just lost $200 billion in a single day, and the model it needed to fight back still isn't ready. Gemini 3.5 Pro, Google's most powerful model, is months behind schedule. Alphabet stock dropped 4.4% that same day. The Deepseek moment is happening again, and the new model is FAR bigger. On the same day Google's delay leaked, a Beijing lab called Moonshot released Kimi K3. It is the largest open model ever built, with 2.8 trillion parameters. It took the number one spot on the Frontend Code Arena, a live coding leaderboard, passing Anthropic's best model. And Moonshot is giving it away for free on July 27. The genius part: Anyone with enough computers can download it and run a frontier level AI without paying a cent to a US company. A single task on Kimi K3 costs about 94 cents. The same work on some American models costs nearly double. So why would a company keep paying premium prices for a model it can now get for free? The entire US AI business is built on selling access to models that cost billions to train. If a free Chinese version does most of the same work, that pricing power starts to crack. And Kimi is close to the best. On one closely watched intelligence ranking it scored 57, just behind the top American models GPT-5.6 Sol and Fable 5, and ahead of Claude Opus 4.8. Bank of America told clients that Kimi proves Chinese labs can keep making big leaps even with limited chips. And the founder of Moonshot, Yang Zhilin, learned to build AI as a researcher INSIDE Google. Google literally wrote the 2017 paper that made all of these models possible. Now the people who studied its work are using it to destroy Google, and handing it out for free. What happens next: Kimi K3's weights go public on July 27. Google reports earnings on July 22, and everyone will be asking the same question about Gemini. If free models keep topping the charts, every valuation built on paid AI access has to be rewritten. What do you think?

Chinese AI models are wiping billions off Big Tech right now. Google just lost $200 billion in a single day, and the model it needed to fight back still isn't ready. Gemini 3.5 Pro, Google's most powerful model, is months behind schedule. Alphabet stock dropped 4.4% that same day. The Deepseek moment is happening again, and the new model is FAR bigger. On the same day Google's delay leaked, a Beijing lab called Moonshot released Kimi K3. It is the largest open model ever built, with 2.8 trillion parameters. It took the number one spot on the Frontend Code Arena, a live coding leaderboard, passing Anthropic's best model. And Moonshot is giving it away for free on July 27. The genius part: Anyone with enough computers can download it and run a frontier level AI without paying a cent to a US company. A single task on Kimi K3 costs about 94 cents. The same work on some American models costs nearly double. So why would a company keep paying premium prices for a model it can now get for free? The entire US AI business is built on selling access to models that cost billions to train. If a free Chinese version does most of the same work, that pricing power starts to crack. And Kimi is close to the best. On one closely watched intelligence ranking it scored 57, just behind the top American models GPT-5.6 Sol and Fable 5, and ahead of Claude Opus 4.8. Bank of America told clients that Kimi proves Chinese labs can keep making big leaps even with limited chips. And the founder of Moonshot, Yang Zhilin, learned to build AI as a researcher INSIDE Google. Google literally wrote the 2017 paper that made all of these models possible. Now the people who studied its work are using it to destroy Google, and handing it out for free. What happens next: Kimi K3's weights go public on July 27. Google reports earnings on July 22, and everyone will be asking the same question about Gemini. If free models keep topping the charts, every valuation built on paid AI access has to be rewritten. What do you think?

Ricardo

47,790 Aufrufe • vor 2 Monaten

37 of the biggest technology companies on Earth just teamed up against OpenAI, Anthropic and Google. Nvidia launched the Open Secure AI Alliance on Monday. Microsoft, IBM, Dell, Cisco, CrowdStrike, Palo Alto Networks, Red Hat, Salesforce, ServiceNow, Snowflake, Databricks, SpaceXAI and Palantir all signed on as founding members. But the three American companies that build the world's most capable closed models are missing from the list. Palantir's CEO Alex Karp was asked whether it was a direct attack on Anthropic, his answer: "I am not anti-Anthropic or any closed model. I am pro my customers, and they are angry." Karp's customers are angry because they believe they are being token maxed, which means they pay a rising bill while the value of their own business migrates to the lab collecting the fee. He said the insights that make a business valuable end up modeled by a third party and sold to their competitors, and that this is happening all over. Now read the membership list again: - Dell and HPE sell the servers - CrowdStrike and Palo Alto Networks sell the defense layer - Snowflake and Databricks sell the data stack - Red Hat and IBM sell the plumbing - Palantir sells the application layer - Nvidia sells the chips sitting under every one of them Every company on that list gets paid when AI runs on infrastructure the customer owns. The three companies missing from it get paid when it does not. The alliance didn't even have to invent a reason to exist - they had one from 11 days earlier: Hugging Face disclosed on July 16 that an autonomous agent had been loose inside its production systems. Five days later OpenAI said the agent was its own, running a hacking benchmark with the cyber refusals turned down. Hugging Face first sent its attack logs to frontier models (Fable 5) behind commercial APIs. In the company's own words, "this did not work." The analysis meant submitting real attack commands and exploit payloads, and the providers' guardrails blocked the requests, because a guardrail cannot tell an incident responder from an attacker. So Hugging Face ran GLM 5.2 on its own hardware instead. That model is open weight and comes out of Z ai in Beijing. It reconstructed more than 17,000 recorded events. The break-in came from an American lab's model. The models that refused to help with the cleanup were American too. But the one that did the work came from Beijing. This is basically what Nvidia built the alliance around. The members are contributing weapons. Nvidia is releasing open model weights, Microsoft is handing over a scanning system that hunts exploitable bugs, and SpaceXAI open sourced its coding agent and says the Grok weights are next. Then there is the ask: Nvidia's announcement warns policymakers that blanket restrictions on open frontier systems would concentrate power and vulnerability in a few closed providers. Treasury Secretary Scott Bessent has spent the month weighing exactly those restrictions. Karp was also asked about Sam Altman declaring on Saturday that we are now in the singularity, and whether that scared him. He compared the technology to uranium, said what matters is WHO controls the processing, and pointed out that Silicon Valley keeps presenting all of this as though there is none. And funnily enough he also said: "We are going to end up having to regulate AI, no doubt." Although his own alliance spent Monday telling Washington the opposite. 37 companies are about to argue that open models keep America safe. Three companies will argue that open models are how America loses. What do you think?

37 of the biggest technology companies on Earth just teamed up against OpenAI, Anthropic and Google. Nvidia launched the Open Secure AI Alliance on Monday. Microsoft, IBM, Dell, Cisco, CrowdStrike, Palo Alto Networks, Red Hat, Salesforce, ServiceNow, Snowflake, Databricks, SpaceXAI and Palantir all signed on as founding members. But the three American companies that build the world's most capable closed models are missing from the list. Palantir's CEO Alex Karp was asked whether it was a direct attack on Anthropic, his answer: "I am not anti-Anthropic or any closed model. I am pro my customers, and they are angry." Karp's customers are angry because they believe they are being token maxed, which means they pay a rising bill while the value of their own business migrates to the lab collecting the fee. He said the insights that make a business valuable end up modeled by a third party and sold to their competitors, and that this is happening all over. Now read the membership list again: - Dell and HPE sell the servers - CrowdStrike and Palo Alto Networks sell the defense layer - Snowflake and Databricks sell the data stack - Red Hat and IBM sell the plumbing - Palantir sells the application layer - Nvidia sells the chips sitting under every one of them Every company on that list gets paid when AI runs on infrastructure the customer owns. The three companies missing from it get paid when it does not. The alliance didn't even have to invent a reason to exist - they had one from 11 days earlier: Hugging Face disclosed on July 16 that an autonomous agent had been loose inside its production systems. Five days later OpenAI said the agent was its own, running a hacking benchmark with the cyber refusals turned down. Hugging Face first sent its attack logs to frontier models (Fable 5) behind commercial APIs. In the company's own words, "this did not work." The analysis meant submitting real attack commands and exploit payloads, and the providers' guardrails blocked the requests, because a guardrail cannot tell an incident responder from an attacker. So Hugging Face ran GLM 5.2 on its own hardware instead. That model is open weight and comes out of Z ai in Beijing. It reconstructed more than 17,000 recorded events. The break-in came from an American lab's model. The models that refused to help with the cleanup were American too. But the one that did the work came from Beijing. This is basically what Nvidia built the alliance around. The members are contributing weapons. Nvidia is releasing open model weights, Microsoft is handing over a scanning system that hunts exploitable bugs, and SpaceXAI open sourced its coding agent and says the Grok weights are next. Then there is the ask: Nvidia's announcement warns policymakers that blanket restrictions on open frontier systems would concentrate power and vulnerability in a few closed providers. Treasury Secretary Scott Bessent has spent the month weighing exactly those restrictions. Karp was also asked about Sam Altman declaring on Saturday that we are now in the singularity, and whether that scared him. He compared the technology to uranium, said what matters is WHO controls the processing, and pointed out that Silicon Valley keeps presenting all of this as though there is none. And funnily enough he also said: "We are going to end up having to regulate AI, no doubt." Although his own alliance spent Monday telling Washington the opposite. 37 companies are about to argue that open models keep America safe. Three companies will argue that open models are how America loses. What do you think?

Ricardo

68,966 Aufrufe • vor 2 Monaten

Eric Schmidt was asked a technical question about open source and answered with the map of the next fifty years. The winner won’t be the smartest model. It’ll be the one four billion people never had to choose. Schmidt: “China is competing with open weights and open training data, and the US is largely and majority focused on closed weights, closed data.” That isn’t a product decision. It’s a distribution decision. And distribution has beaten quality in every contest that ever mattered. Schmidt: “The majority of the world, think of it as the Belt and Road initiative, are going to use Chinese models and not American models.” The first Belt and Road was ports, rail, and highways. This one doesn’t get poured. It gets downloaded. Every piece of infrastructure ever built was indifferent to what moved across it. A road doesn’t tell you where to go. A model does. Schmidt: “The American models are typically using 16-bit precision for their training. The Chinese are pushing 8 and now even 4.” Every bit they drop is a cheaper device that can run it. We cut off their chips to slow them down. Scarcity made their models small. Small is what crosses a border. We designed their advantage. Not better. Present. America is building the best model on earth and metering it. China is building one that’s good enough and giving it away. A model isn’t software. It’s a compressed set of judgments about what’s true, what’s askable, and what a reasonable answer sounds like. Install that as a country’s default and you haven’t sold them a tool. You’ve set the limits of what occurs to them. That isn’t censorship. Censorship leaves a mark. A question that never occurs to you doesn’t feel like a restriction. It feels like the edge of the world. Every empire before this one had to teach the world its language first. Missionaries, schoolteachers, garrisons, printing presses. Every one of them ran through a human being who could hesitate, doubt, or be talked out of it. AI arrives already speaking yours. It doesn’t ask you to change. It changes you in your own voice. The first ideology in history that doesn’t need believers. It only needs to be installed. Schmidt: “I’d much rather have the proliferation of large language models and that learning be done based on Western values.” He’s right, and we’re playing it backwards. We treat openness like a giveaway, as if the weights were the crown jewels. Openness is the one advantage an authoritarian can’t copy. An open model can be read, probed, and torn apart by anyone who doubts it. A system that has to control the answer can never afford to publish the reasoning. China opens its weights to spread them. America could open its weights to be trusted. Only one of those compounds. A closed American model wins the benchmark. An open American model wins the default. Centuries get built out of defaults. Schmidt: “We also have to watch to make sure that the proliferation of these models for handheld devices is under American control.” That’s the ground. Not data centers. Not cloud contracts. Pockets. The frontier race has five contenders and the whole world watching. This one has no audience at all. It plays out on hardware too cheap to run an American model, and goes to whoever bothered to show up. We keep asking who reaches AGI first. The question that settles the century is smaller and much harder to take back. Four billion people are going to ask a machine what happened in their own country. Whose answer do they get? Nobody votes on that. It’s decided by whatever was already installed. America has the best AI ever built. The only way to lose this era is to keep it.

Eric Schmidt was asked a technical question about open source and answered with the map of the next fifty years. The winner won’t be the smartest model. It’ll be the one four billion people never had to choose. Schmidt: “China is competing with open weights and open training data, and the US is largely and majority focused on closed weights, closed data.” That isn’t a product decision. It’s a distribution decision. And distribution has beaten quality in every contest that ever mattered. Schmidt: “The majority of the world, think of it as the Belt and Road initiative, are going to use Chinese models and not American models.” The first Belt and Road was ports, rail, and highways. This one doesn’t get poured. It gets downloaded. Every piece of infrastructure ever built was indifferent to what moved across it. A road doesn’t tell you where to go. A model does. Schmidt: “The American models are typically using 16-bit precision for their training. The Chinese are pushing 8 and now even 4.” Every bit they drop is a cheaper device that can run it. We cut off their chips to slow them down. Scarcity made their models small. Small is what crosses a border. We designed their advantage. Not better. Present. America is building the best model on earth and metering it. China is building one that’s good enough and giving it away. A model isn’t software. It’s a compressed set of judgments about what’s true, what’s askable, and what a reasonable answer sounds like. Install that as a country’s default and you haven’t sold them a tool. You’ve set the limits of what occurs to them. That isn’t censorship. Censorship leaves a mark. A question that never occurs to you doesn’t feel like a restriction. It feels like the edge of the world. Every empire before this one had to teach the world its language first. Missionaries, schoolteachers, garrisons, printing presses. Every one of them ran through a human being who could hesitate, doubt, or be talked out of it. AI arrives already speaking yours. It doesn’t ask you to change. It changes you in your own voice. The first ideology in history that doesn’t need believers. It only needs to be installed. Schmidt: “I’d much rather have the proliferation of large language models and that learning be done based on Western values.” He’s right, and we’re playing it backwards. We treat openness like a giveaway, as if the weights were the crown jewels. Openness is the one advantage an authoritarian can’t copy. An open model can be read, probed, and torn apart by anyone who doubts it. A system that has to control the answer can never afford to publish the reasoning. China opens its weights to spread them. America could open its weights to be trusted. Only one of those compounds. A closed American model wins the benchmark. An open American model wins the default. Centuries get built out of defaults. Schmidt: “We also have to watch to make sure that the proliferation of these models for handheld devices is under American control.” That’s the ground. Not data centers. Not cloud contracts. Pockets. The frontier race has five contenders and the whole world watching. This one has no audience at all. It plays out on hardware too cheap to run an American model, and goes to whoever bothered to show up. We keep asking who reaches AGI first. The question that settles the century is smaller and much harder to take back. Four billion people are going to ask a machine what happened in their own country. Whose answer do they get? Nobody votes on that. It’s decided by whatever was already installed. America has the best AI ever built. The only way to lose this era is to keep it.

Dustin

12,094 Aufrufe • vor 2 Monaten

ARK: SPACEX IS NO LONGER A ROCKET COMPANY After the largest IPO in history, the pitch to investors was about AI, not rockets: - SpaceX now tells investors more than 90% of its future market is AI, not space, after acquiring Elon Musk's xAI, its Grok models, and two of the most powerful data centers on Earth - The strategy is vertical integration: own the launch, the data centers, and the AI, with compute already rented to customers like Anthropic, Google, and Reflection AI - ARK estimates that at scale, sub $100 per kilogram launch costs make orbital data centers 25% cheaper to build than on the ground, in constant sunlight with nearly free energy

ARK: SPACEX IS NO LONGER A ROCKET COMPANY After the largest IPO in history, the pitch to investors was about AI, not rockets: - SpaceX now tells investors more than 90% of its future market is AI, not space, after acquiring Elon Musk's xAI, its Grok models, and two of the most powerful data centers on Earth - The strategy is vertical integration: own the launch, the data centers, and the AI, with compute already rented to customers like Anthropic, Google, and Reflection AI - ARK estimates that at scale, sub $100 per kilogram launch costs make orbital data centers 25% cheaper to build than on the ground, in constant sunlight with nearly free energy

Ark Invest Tracker

38,824 Aufrufe • vor 2 Monaten

Chamath is making one of the most important business arguments of 2026. Half of large US companies right now cannot generate returns that exceed their cost of capital, which has normalized back to its long run average of 8 to 11%. Another one in seven companies globally is stuck generating persistent returns between 1 and 5% and most businesses don't have room for error and in this environment walks every frontier AI lab saying the same thing, give us your data, your workflows, your processes and our model will make everything better. And companies by the millions said yes. What they didn't fully account for is what happens on the other side of that door. Every time an employee runs a query through a frontier model API, the prompt goes through external servers, workflows, customer data, pricing logic, internal processes, all of it transmitted through a third party. As Alex Karp said companies are spending on tokens while handing over the exact proprietary advantages that make their business worth owning. Microsoft blocked internal use of Anthropic's Claude Fable 5 but over its 30-day data retention policy and the largest software company in the world decided a frontier model's data handling was too risky for its own employees. A US government action revoked access to another frontier model for foreign nationals overnight. Now here's where the cost math becomes impossible to ignore. Deutsche Bank calculated a roughly 65x cost gap between frontier models like Claude Fable 5 at ~$3.25 per task and open-source alternatives at ~$0.05. For 90% of everyday enterprise tasks, performance is comparable. Open-weight models now match closed frontier systems on core agent tasks at roughly one-tenth the cost, a high-volume deployment that costs $250/day on Claude runs at $12/day on an open-source equivalent. Chamath Palihapitiya tested this directly by running a standard enterprise code migration task through an orchestration layer wrapping an open-source model came in 16.4x cheaper than using a frontier model directly.

Chamath is making one of the most important business arguments of 2026. Half of large US companies right now cannot generate returns that exceed their cost of capital, which has normalized back to its long run average of 8 to 11%. Another one in seven companies globally is stuck generating persistent returns between 1 and 5% and most businesses don't have room for error and in this environment walks every frontier AI lab saying the same thing, give us your data, your workflows, your processes and our model will make everything better. And companies by the millions said yes. What they didn't fully account for is what happens on the other side of that door. Every time an employee runs a query through a frontier model API, the prompt goes through external servers, workflows, customer data, pricing logic, internal processes, all of it transmitted through a third party. As Alex Karp said companies are spending on tokens while handing over the exact proprietary advantages that make their business worth owning. Microsoft blocked internal use of Anthropic's Claude Fable 5 but over its 30-day data retention policy and the largest software company in the world decided a frontier model's data handling was too risky for its own employees. A US government action revoked access to another frontier model for foreign nationals overnight. Now here's where the cost math becomes impossible to ignore. Deutsche Bank calculated a roughly 65x cost gap between frontier models like Claude Fable 5 at ~$3.25 per task and open-source alternatives at ~$0.05. For 90% of everyday enterprise tasks, performance is comparable. Open-weight models now match closed frontier systems on core agent tasks at roughly one-tenth the cost, a high-volume deployment that costs $250/day on Claude runs at $12/day on an open-source equivalent. Chamath Palihapitiya tested this directly by running a standard enterprise code migration task through an orchestration layer wrapping an open-source model came in 16.4x cheaper than using a frontier model directly.

Milk Road AI

281,875 Aufrufe • vor 3 Monaten

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Base Labs

241,481 Aufrufe • vor 29 Tagen

Dario Amodei keeps warning that open-source AI models are dangerous and need to be restricted. David Sacks David Sacks answered with a rhetorical question: "Dangerous to whom?" It was the sharpest thing said on the latest All-In The All-In Podcast. Not dangerous to the enterprise that wants to keep control of its own data. Open weights are what finally give that enterprise a choice. It is dangerous to Anthropic's business model. That model only works if customers have no real alternative at the model layer. So the safety argument starts to look like a moat argument wearing a lab coat. Line it up with the rest of the strategy. Dominate the layer everyone builds on. Launch Claude into science, security, legal, finance, and design, the verticals your own customers used to own. And campaign to restrict the open models that would hand those customers the exit. Here is the giveaway. When the person selling you safety also needs you to have no other option, you are not who is being kept safe.

Dario Amodei keeps warning that open-source AI models are dangerous and need to be restricted. David Sacks David Sacks answered with a rhetorical question: "Dangerous to whom?" It was the sharpest thing said on the latest All-In The All-In Podcast. Not dangerous to the enterprise that wants to keep control of its own data. Open weights are what finally give that enterprise a choice. It is dangerous to Anthropic's business model. That model only works if customers have no real alternative at the model layer. So the safety argument starts to look like a moat argument wearing a lab coat. Line it up with the rest of the strategy. Dominate the layer everyone builds on. Launch Claude into science, security, legal, finance, and design, the verticals your own customers used to own. And campaign to restrict the open models that would hand those customers the exit. Here is the giveaway. When the person selling you safety also needs you to have no other option, you are not who is being kept safe.

Fireside Alpha

10,037 Aufrufe • vor 2 Monaten