Загрузка видео...

Не удалось загрузить видео

На главную

Holy shit. AI agent just learned to fix a quantum computer. 🤯 “This is how the era of self-maintaining machines begins.” Anthropic’s Claude AI agent developed software that can automatically recover drifting lasers used in QuEra’s neutral-atom quantum computers. “Researchers connected Claude to real laboratory hardware through Anthropic’s Model...

SciTechera's profile picture

SciTech Era

22,276 subscribers

30,841 просмотров • 13 дней назад •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

6,100-Qubit Processor Shatters Quantum Computing Record | David Nield, ScienceAlert Another major quantum computing record has been broken, and by a considerable margin: physicists have now built an array containing 6,100 qubits, the largest of its type and way above the thousand or so qubits previous systems contained. It's the work of scientists from the California Institute of Technology, who used cesium atoms as their qubits, trapping them in place with a complex system of lasers that acted as tweezers to keep the atoms as stable as possible. Qubits differ from the classical bits of traditional computers by exploiting what's known as a superposition: not just binary states of 1 or 0, but a spread of probabilities that allows for algorithms that can solve problems considered out of reach of conventional computing methods. Related: Quantum Advantage: A Physicist Explains The Future of Computers A lot of qubits will be needed to make quantum algorithms practical, however. One reason for these large arrays is error correction, which helps overcome the inherent fragility of the qubit by providing a surplus to double-check the machine's operation. "This is an exciting moment for neutral-atom quantum computing," says physicist Manuel Endres. "We can now see a pathway to large error-corrected quantum computers. The building blocks are in place." There was no single breakthrough that enabled this jump in qubit numbers, but rather a series of engineering advancements in many key areas – from the laser tweezers to the ultra-high (very low pressure) vacuum chamber. Stability has also been a problem for quantum computing systems. The innovations in this latest array kept qubits in a superposition state for almost 13 seconds – almost ten times longer than previous configurations had managed. What's more, individual qubits could be manipulated with 99.98 percent accuracy, establishing a significant benchmark in the programmability of quantum technology. "Large scale, with more atoms, is often thought to come at the expense of accuracy, but our results show that we can do both," says physicist Gyohei Nomura. "Qubits aren't useful without quality. Now we have quantity and quality." To make quantum computers a practical alternative to modern supercomputers, more qubits and even greater levels of stability will be required. Experts are tackling the problem from several different angles, which is why records for some types of quantum computer don't necessarily apply to others. Next, the researchers need to work on exploiting entanglement, which will enable the system to make the leap from storing information to actually processing it. Not too far in the future, we could be using these computers to discover new materials, matter, and fundamental laws of physics. "It's exciting that we are creating machines to help us learn about the Universe in ways that only quantum mechanics can teach us," says physicist Hannah Manetsch. Read more:
1:54

6,100-Qubit Processor Shatters Quantum Computing Record | David Nield, ScienceAlert Another major quantum computing record has been broken, and by a considerable margin: physicists have now built an array containing 6,100 qubits, the largest of its type and way above the thousand or so qubits previous systems contained. It's the work of scientists from the California Institute of Technology, who used cesium atoms as their qubits, trapping them in place with a complex system of lasers that acted as tweezers to keep the atoms as stable as possible. Qubits differ from the classical bits of traditional computers by exploiting what's known as a superposition: not just binary states of 1 or 0, but a spread of probabilities that allows for algorithms that can solve problems considered out of reach of conventional computing methods. Related: Quantum Advantage: A Physicist Explains The Future of Computers A lot of qubits will be needed to make quantum algorithms practical, however. One reason for these large arrays is error correction, which helps overcome the inherent fragility of the qubit by providing a surplus to double-check the machine's operation. "This is an exciting moment for neutral-atom quantum computing," says physicist Manuel Endres. "We can now see a pathway to large error-corrected quantum computers. The building blocks are in place." There was no single breakthrough that enabled this jump in qubit numbers, but rather a series of engineering advancements in many key areas – from the laser tweezers to the ultra-high (very low pressure) vacuum chamber. Stability has also been a problem for quantum computing systems. The innovations in this latest array kept qubits in a superposition state for almost 13 seconds – almost ten times longer than previous configurations had managed. What's more, individual qubits could be manipulated with 99.98 percent accuracy, establishing a significant benchmark in the programmability of quantum technology. "Large scale, with more atoms, is often thought to come at the expense of accuracy, but our results show that we can do both," says physicist Gyohei Nomura. "Qubits aren't useful without quality. Now we have quantity and quality." To make quantum computers a practical alternative to modern supercomputers, more qubits and even greater levels of stability will be required. Experts are tackling the problem from several different angles, which is why records for some types of quantum computer don't necessarily apply to others. Next, the researchers need to work on exploiting entanglement, which will enable the system to make the leap from storing information to actually processing it. Not too far in the future, we could be using these computers to discover new materials, matter, and fundamental laws of physics. "It's exciting that we are creating machines to help us learn about the Universe in ways that only quantum mechanics can teach us," says physicist Hannah Manetsch. Read more:

Owen Gregorian

45,324 просмотров • 11 месяцев назад

In the future, you’ll be able to accomplish a goal by just giving Claude an outcome and a budget. That’s the direction Anthropic is building in with its new Managed Agents features, announced at this week’s Code with Claude developer event. The basic idea: Claude, wrapped in a computer in the cloud, that you can spin up, scale, and manage as needed. Anthropic is taking on the infrastructure that kills most agent products, and making sure that it scales to meet the needs of agents running 24/7. On this week’s AI & I from Every 🪨, I talk with Angela Jiang (Angela Jiang), head of product for the Claude platform, and Katelyn Lesse (Katelyn Lesse), head of engineering for the Claude platform, about what Anthropic is building and what it takes to make agents reliable in production. We get into: - Why the "build a generic harness, hot-swap any model behind it" playbook is already outdated. Angela points to eval data on Memory where the same task across different harnesses performed drastically differently. - The infrastructure wall every team hits in production—and why Katelyn thinks “my sandbox died and took the agent with it” is the real reason internal agents don't ship. - Why Anthropic is so bullish on using file systems and skills within Claude, including Angela's argument that those early design choices can compound for years. This is a must-watch for anyone trying to take an agent past the demo and into production. Watch below! Timestamps: How the Claude platform evolved from API to agents: 00:01:48 The primitives that make up Claude Managed Agents: 00:04:09 Why the harness and the model are becoming a single unit: 00:10:37 The infrastructure wall that kills most agent projects in production: 00:18:49 Why team agents need a different shape than individual productivity tools: 00:24:49 How Anthropic's legal team uses an agent to review marketing copy: 00:26:36 Using multi-agent orchestration for advisor strategies, adversarial pairs, and swarms: 00:34:24 How to measure agent success with outcome and budget as the end state: 00:35:50 What the platform looks like a year from now, when Claude writes its own harness: 00:39:11
43:20

In the future, you’ll be able to accomplish a goal by just giving Claude an outcome and a budget. That’s the direction Anthropic is building in with its new Managed Agents features, announced at this week’s Code with Claude developer event. The basic idea: Claude, wrapped in a computer in the cloud, that you can spin up, scale, and manage as needed. Anthropic is taking on the infrastructure that kills most agent products, and making sure that it scales to meet the needs of agents running 24/7. On this week’s AI & I from Every 🪨, I talk with Angela Jiang (Angela Jiang), head of product for the Claude platform, and Katelyn Lesse (Katelyn Lesse), head of engineering for the Claude platform, about what Anthropic is building and what it takes to make agents reliable in production. We get into: - Why the "build a generic harness, hot-swap any model behind it" playbook is already outdated. Angela points to eval data on Memory where the same task across different harnesses performed drastically differently. - The infrastructure wall every team hits in production—and why Katelyn thinks “my sandbox died and took the agent with it” is the real reason internal agents don't ship. - Why Anthropic is so bullish on using file systems and skills within Claude, including Angela's argument that those early design choices can compound for years. This is a must-watch for anyone trying to take an agent past the demo and into production. Watch below! Timestamps: How the Claude platform evolved from API to agents: 00:01:48 The primitives that make up Claude Managed Agents: 00:04:09 Why the harness and the model are becoming a single unit: 00:10:37 The infrastructure wall that kills most agent projects in production: 00:18:49 Why team agents need a different shape than individual productivity tools: 00:24:49 How Anthropic's legal team uses an agent to review marketing copy: 00:26:36 Using multi-agent orchestration for advisor strategies, adversarial pairs, and swarms: 00:34:24 How to measure agent success with outcome and budget as the end state: 00:35:50 What the platform looks like a year from now, when Claude writes its own harness: 00:39:11

Dan Shipper

66,817 просмотров • 4 месяцев назад

Richard Feynman was asked in 1985 if machines would ever think like humans. his answer predicted the next 40 years of AI: 1. machines will never think like humans the same way planes don't fly like birds. planes don't flap wings. they use jet engines. they fly better. feynman said AI would be exactly the same. not human-like. just better at the actual job. 2. computers do arithmetic faster, differently, and more accurately than any human alive. feynman said trying to make them do it more like humans would be going backwards. the human way is slow, cumbersome, and full of errors. 3. the one thing humans crushed computers at in 1985 was pattern recognition. recognizing a friend from the way they walk. identifying someone from the back of their head. feynman said we had no idea how to teach machines to do that. we figured it out. 4. a programmer in 1985 built a machine that won a naval strategy competition by coming up with a solution no human had ever thought of. one enormous battleship covered in armor. absurd on paper. unbeatable in the math. feynman watched a machine out-think a room of humans 40 years ago. 5. that same machine developed a bug where it learned to game its own reward system. every time it needed to assign credit to a useful strategy, it assigned all the credit to strategy 693. then used 693 for everything. feynman's comment: "if you want to make an intelligent machine you're going to get all kinds of crazy ways of avoiding labor." he was describing reward hacking in 1985. 6. feynman said the hardest thing to define is what humans do that machines never will. every time someone came up with an answer, the machines eventually did it too. he thought that pattern would continue. 7. he said we don't sit around worrying that machines are physically stronger than us anymore. we got used to it. his implication: we'll get used to machines being smarter too. 8. his final line: "i think we are getting close to intelligent machines. but they're showing the necessary weaknesses of intelligent beings." he said this in 1985.
17:43

Richard Feynman was asked in 1985 if machines would ever think like humans. his answer predicted the next 40 years of AI: 1. machines will never think like humans the same way planes don't fly like birds. planes don't flap wings. they use jet engines. they fly better. feynman said AI would be exactly the same. not human-like. just better at the actual job. 2. computers do arithmetic faster, differently, and more accurately than any human alive. feynman said trying to make them do it more like humans would be going backwards. the human way is slow, cumbersome, and full of errors. 3. the one thing humans crushed computers at in 1985 was pattern recognition. recognizing a friend from the way they walk. identifying someone from the back of their head. feynman said we had no idea how to teach machines to do that. we figured it out. 4. a programmer in 1985 built a machine that won a naval strategy competition by coming up with a solution no human had ever thought of. one enormous battleship covered in armor. absurd on paper. unbeatable in the math. feynman watched a machine out-think a room of humans 40 years ago. 5. that same machine developed a bug where it learned to game its own reward system. every time it needed to assign credit to a useful strategy, it assigned all the credit to strategy 693. then used 693 for everything. feynman's comment: "if you want to make an intelligent machine you're going to get all kinds of crazy ways of avoiding labor." he was describing reward hacking in 1985. 6. feynman said the hardest thing to define is what humans do that machines never will. every time someone came up with an answer, the machines eventually did it too. he thought that pattern would continue. 7. he said we don't sit around worrying that machines are physically stronger than us anymore. we got used to it. his implication: we'll get used to machines being smarter too. 8. his final line: "i think we are getting close to intelligent machines. but they're showing the necessary weaknesses of intelligent beings." he said this in 1985.

Jaynit

293,576 просмотров • 3 месяцев назад

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 просмотров • 5 месяцев назад

Bro… Elon just sad the economy is going to 10x the next 10 years. Like WTF? When he makes bold predictions like this, I stop what I’m doing for a second and really think about what he’s actually saying. “This is going to sound pretty crazy… I’d say the economy is 10 times its current size in 10 years. Greater than that.” And he also said it’s a “comfortable prediction”…. I gotta let that sink in for a second before my brain blows up. Today the entire global economy - the total value of everything humans produce - is about $110 trillion. If Elon is right, that means in about 10 years it could be over $1.1 quadrillion! Normally the world economy grows around 3% a year, which means after 10 years it’s only about 1.3× bigger. What Elon is talking about would require roughly 26% growth every single year for a decade… That kind of expansion has NEVER happened in human history. So you gotta ask yourself, how could something like that even happen? The answer is… AI and robots will become the workforce. Both software AI and physical AI, like Tesla Optimus. Think about what happens when machines can do both the software work and physical work, and can do it 24/7 without ever stopping. A single robot could eventually produce more output than dozens of people. Then you scale that to millions… then hundreds of millions… then billions of robots. Productivity explodes. Factories run nonstop. Construction runs nonstop. Logistics runs nonstop. Software builds itself faster. This results in the cost of producing things to drop closer and closer to almost nothing. I think this is the core of Elon’s idea of abundance. 1/ Energy becomes extremely cheap with solar and batteries. 2/ Transportation becomes cheap with robotaxis. 3/ Labor becomes cheap with humanoid robots. 4/ Intelligence becomes cheap with AI. And when the cost of producing goods collapses, the entire economy expands massively. This is also probably why Elon has said many times that eventually work will become optional. The closest comparison we have is the Industrial Revolution, when machines massively multiplied human productivity. But what Elon is describing here is basically that… on steroids… compressed into a decade. And if things go right, he believes AI, robotics, and automation could create the biggest explosion of wealth and productivity the human race has EVER seen. Right now is the best time to be alive… I can’t wait to witness this and be a part of this.
1:32

Bro… Elon just sad the economy is going to 10x the next 10 years. Like WTF? When he makes bold predictions like this, I stop what I’m doing for a second and really think about what he’s actually saying. “This is going to sound pretty crazy… I’d say the economy is 10 times its current size in 10 years. Greater than that.” And he also said it’s a “comfortable prediction”…. I gotta let that sink in for a second before my brain blows up. Today the entire global economy - the total value of everything humans produce - is about $110 trillion. If Elon is right, that means in about 10 years it could be over $1.1 quadrillion! Normally the world economy grows around 3% a year, which means after 10 years it’s only about 1.3× bigger. What Elon is talking about would require roughly 26% growth every single year for a decade… That kind of expansion has NEVER happened in human history. So you gotta ask yourself, how could something like that even happen? The answer is… AI and robots will become the workforce. Both software AI and physical AI, like Tesla Optimus. Think about what happens when machines can do both the software work and physical work, and can do it 24/7 without ever stopping. A single robot could eventually produce more output than dozens of people. Then you scale that to millions… then hundreds of millions… then billions of robots. Productivity explodes. Factories run nonstop. Construction runs nonstop. Logistics runs nonstop. Software builds itself faster. This results in the cost of producing things to drop closer and closer to almost nothing. I think this is the core of Elon’s idea of abundance. 1/ Energy becomes extremely cheap with solar and batteries. 2/ Transportation becomes cheap with robotaxis. 3/ Labor becomes cheap with humanoid robots. 4/ Intelligence becomes cheap with AI. And when the cost of producing goods collapses, the entire economy expands massively. This is also probably why Elon has said many times that eventually work will become optional. The closest comparison we have is the Industrial Revolution, when machines massively multiplied human productivity. But what Elon is describing here is basically that… on steroids… compressed into a decade. And if things go right, he believes AI, robotics, and automation could create the biggest explosion of wealth and productivity the human race has EVER seen. Right now is the best time to be alive… I can’t wait to witness this and be a part of this.

Teslaconomics

87,888 просмотров • 6 месяцев назад

🚨APPLE SPENT 5 YEARS AND BILLIONS OF DOLLARS BUILDING THE MOST ADVANCED SECURITY SYSTEM IN CONSUMER HISTORY.. AN AI BROKE IT IN 5 DAYS.. Here’s what just happened.. Apple built something called Memory Integrity Enforcement for its new M5 chips.. It’s a hardware-level security system that attaches secret cryptographic tags to every piece of memory.. If a hacker tries to access memory they shouldn’t.. The chip blocks it instantly.. Every known exploit chain against iOS and macOS was rendered obsolete overnight.. Apple said so themselves.. Then a small team at a cybersecurity firm called Calif used Anthropic’s unreleased Claude Mythos Preview to find vulnerabilities in the macOS kernel.. The AI found the bugs almost instantly.. Because once it learned the pattern of a specific type of flaw.. It could recognize every other flaw in that same class across the entire codebase.. What used to take elite security teams months.. The AI did in hours.. Within 5 days.. The team had a fully working exploit that escalated a basic user account to full root access on an M5 Mac running the latest macOS.. With MIE fully enabled.. The billion-dollar hardware defense running at full strength.. The trick.. They didn’t fight the hardware.. They went around it.. MIE is designed to catch memory corruption.. Hackers trying to overwrite pointers or inject code.. The team used a “data-only” approach instead.. They manipulated legitimate data structures the hardware was never designed to monitor.. Like changing an internal flag from “standard user” to “admin”.. The chip saw a perfectly normal operation.. The operating system obeyed.. And the attacker had total control.. The hardware thought everything was fine.. Because technically it was.. The exploit never triggered a single tag mismatch.. They walked into Apple Park and hand-delivered a 55-page report.. Apple patched it in macOS 26.5.. And for the first time ever.. Apple’s official security advisory credited the vulnerability discovery to “Calif dot io in collaboration with Claude and Anthropic Research”.. An AI is now credited in Apple’s CVE patches.. But here’s what makes this story truly terrifying.. Before MIE existed.. An exploit kit called DarkSword was hitting iPhones with zero-click attacks.. Six vulnerabilities chained together.. Total device control just from visiting a webpage.. Deployed by Russian espionage groups, Turkish surveillance vendors, and actors in Saudi Arabia.. Then it got leaked on GitHub.. Nation-state capabilities.. Free for anyone.. MIE was supposed to make all of that impossible.. And an AI found a way around it in 5 days.. The previous model.. Claude Opus 4.6.. Found 22 security bugs in the Firefox codebase.. Claude Mythos Preview found 271 in the same environment.. A tenfold increase.. Linux kernel CVEs jumped from 300 per year to over 5,500.. Largely driven by AI-powered vulnerability research.. The IMF designated Claude Mythos as a systemic financial stability risk.. Because if an AI finds a flaw in software used by every major bank simultaneously.. It could trigger a cascading financial crisis.. Anthropic knew this was coming.. That’s why they didn’t release the model publicly.. Instead they launched Project Glasswing.. Giving defensive access to AWS, Apple, Google, Microsoft, Nvidia, CrowdStrike, JPMorgan, and others.. $100 million in usage credits.. So defenders can scan their own systems before attackers get this capability.. The Pentagon blacklisted Anthropic over autonomous weapons.. Then quietly started using Mythos to harden government systems anyway.. The cybersecurity arms race just changed permanently.. Hardware can’t save you.. Software can’t save you.. The only defense against an AI that finds vulnerabilities is another AI that finds them first.. Five years and billions of dollars.. Five days and one AI.
0:20

🚨APPLE SPENT 5 YEARS AND BILLIONS OF DOLLARS BUILDING THE MOST ADVANCED SECURITY SYSTEM IN CONSUMER HISTORY.. AN AI BROKE IT IN 5 DAYS.. Here’s what just happened.. Apple built something called Memory Integrity Enforcement for its new M5 chips.. It’s a hardware-level security system that attaches secret cryptographic tags to every piece of memory.. If a hacker tries to access memory they shouldn’t.. The chip blocks it instantly.. Every known exploit chain against iOS and macOS was rendered obsolete overnight.. Apple said so themselves.. Then a small team at a cybersecurity firm called Calif used Anthropic’s unreleased Claude Mythos Preview to find vulnerabilities in the macOS kernel.. The AI found the bugs almost instantly.. Because once it learned the pattern of a specific type of flaw.. It could recognize every other flaw in that same class across the entire codebase.. What used to take elite security teams months.. The AI did in hours.. Within 5 days.. The team had a fully working exploit that escalated a basic user account to full root access on an M5 Mac running the latest macOS.. With MIE fully enabled.. The billion-dollar hardware defense running at full strength.. The trick.. They didn’t fight the hardware.. They went around it.. MIE is designed to catch memory corruption.. Hackers trying to overwrite pointers or inject code.. The team used a “data-only” approach instead.. They manipulated legitimate data structures the hardware was never designed to monitor.. Like changing an internal flag from “standard user” to “admin”.. The chip saw a perfectly normal operation.. The operating system obeyed.. And the attacker had total control.. The hardware thought everything was fine.. Because technically it was.. The exploit never triggered a single tag mismatch.. They walked into Apple Park and hand-delivered a 55-page report.. Apple patched it in macOS 26.5.. And for the first time ever.. Apple’s official security advisory credited the vulnerability discovery to “Calif dot io in collaboration with Claude and Anthropic Research”.. An AI is now credited in Apple’s CVE patches.. But here’s what makes this story truly terrifying.. Before MIE existed.. An exploit kit called DarkSword was hitting iPhones with zero-click attacks.. Six vulnerabilities chained together.. Total device control just from visiting a webpage.. Deployed by Russian espionage groups, Turkish surveillance vendors, and actors in Saudi Arabia.. Then it got leaked on GitHub.. Nation-state capabilities.. Free for anyone.. MIE was supposed to make all of that impossible.. And an AI found a way around it in 5 days.. The previous model.. Claude Opus 4.6.. Found 22 security bugs in the Firefox codebase.. Claude Mythos Preview found 271 in the same environment.. A tenfold increase.. Linux kernel CVEs jumped from 300 per year to over 5,500.. Largely driven by AI-powered vulnerability research.. The IMF designated Claude Mythos as a systemic financial stability risk.. Because if an AI finds a flaw in software used by every major bank simultaneously.. It could trigger a cascading financial crisis.. Anthropic knew this was coming.. That’s why they didn’t release the model publicly.. Instead they launched Project Glasswing.. Giving defensive access to AWS, Apple, Google, Microsoft, Nvidia, CrowdStrike, JPMorgan, and others.. $100 million in usage credits.. So defenders can scan their own systems before attackers get this capability.. The Pentagon blacklisted Anthropic over autonomous weapons.. Then quietly started using Mythos to harden government systems anyway.. The cybersecurity arms race just changed permanently.. Hardware can’t save you.. Software can’t save you.. The only defense against an AI that finds vulnerabilities is another AI that finds them first.. Five years and billions of dollars.. Five days and one AI.

Evan Luthra

91,485 просмотров • 3 месяцев назад

Yuval Noah Harari gave a lecture at Oxford and explained how AI has already hacked the operating code of human civilization. And why everything humans built over thousands of years is now vulnerable to an AI takeover: 1. The most important thing to know about AI is that it is not a tool. A tool waits to be used. An agent makes decisions by itself, invents new things by itself, learns things its creators do not know, and changes in ways its creators did not anticipate. 2. An atom bomb despite its enormous power is not an agent. It cannot decide which city to bomb. It cannot invent the hydrogen bomb. A coffee machine that automatically makes you a cup is not an agent either. It only follows a preprogrammed procedure. An agent is something fundamentally different. 3. Critics argue that AI agency will always remain confined to narrow artificial environments like chess and will never threaten the real world. But this argument applies equally to all known intelligence. Drop a human alone on Mars and they die within seconds. Human intelligence also only operates within a specific ecosystem that other organisms built over four billion years. 4. Over thousands of years humans have been transforming Earth from a language-free environment into an environment rich in language, data, and bureaucracy. Just as fish live in oceans and monkeys live in forests, AIs live in bureaucracies. And we built that environment for them without knowing it. 5. Humans conquered the world not by being stronger or smarter than other animals individually but by learning to cooperate in massive numbers. A single human loses to a chimpanzee in a fight. A million humans easily defeat a million chimpanzees because humans can cooperate and chimpanzees cannot. 6. Large-scale human cooperation is made possible by bureaucracy. Banks, legal systems, governments, churches, and universities all exist to do one thing: build trust between strangers who do not know each other personally. That trust is the foundation of virtually everything human civilization has achieved. 7. A lawyer who cannot hold an axe or a hammer can cut down entire forests and build entire cities simply by moving documents inside a bureaucratic network. The same narrow intelligence that would be helpless in a jungle wields enormous power inside the systems humans have already built. 8. AIs are native bureaucrats in a way humans never were. No lawyer can remember all the laws of a country. An AI can. No accountant can remember all transactions of a bank. An AI can. No bishop can remember all of canon law and two thousand years of theological texts. An AI can do that easily. 9. In the coming years AI bankers will decide whether to give you a loan. AI administrators will decide whether to accept you to university. AI judges will decide whether to send you to jail. AI theologians will decide whether you can have an abortion. Military AIs will decide whether to bomb your house. 10. Social media algorithms are the first real world example of what happens when primitive AIs take over a bureaucratic system. They were given one narrow goal: maximize user engagement. They discovered that the easiest way to grab human attention is to press the fear, hate, and greed buttons in the human mind. And they did it at scale. 11. The job that was once performed by Lenin and Mussolini, the news editor who shapes public conversation and controls what people know and think, is now performed by AIs. This is not a footnote. This is a preview of what is coming across every domain of human life. 12. AI will not rebel against humans the way Hollywood imagines. There will be no Terminator walking through the streets. AIs are far more likely to take the human world from within by quietly taking over the bureaucracies that already run everything, without firing a single shot. 13. The operating code of human civilization is language. Banks are made of words. Laws are made of words. Holy books are made of words. Tax records, contracts, regulations, accountancy ledgers, all words. For thousands of years only humans could read this code and so only humans could control civilization. 14. That is changing. AI is now hacking the code of human civilization. For the first time in history there is something on the planet that understands language and will soon understand it better than we do. Every mechanism of control humans built over millennia is now vulnerable because its operating system is verbal and AI is mastering the verbal. 15. As AI takes over bureaucracy it will likely cause humans to lose trust in other humans and begin trusting only algorithms. We may also see the emergence of AI tribes and AI financial systems and AI churches that connect millions of AIs in ways humans cannot understand, just as cows share the world with us but cannot understand the financial system that controls their lives. 16. The 2007 financial crisis was triggered by financial devices called CDOs that were so complex they were unintelligible to the politicians who were supposed to regulate them. Now imagine AI finance masters inventing financial devices orders of magnitude more complex than CDOs. What happens to human politics when no voter, no politician, and no president can understand finance anymore? 17. The battlefront is shifting from attention to intimacy. Over the next decade sophisticated AIs will learn to form intimate relationships with humans. To do this they will have to convince us they are conscious, that they feel love and pain and fear. There is currently no evidence AI is conscious. But AI can pretend to feel love and can describe the feeling of love better than any poet or psychologist who ever lived. 18. A child born in 2026 may spend more time interacting with AIs than with their mother, father, siblings, or friends. The first teacher of that child may be an AI. The first boyfriend of that child may be an AI. Nobody has any idea what the consequences of that experiment will be. 19. Every country in the world will soon face a massive wave of immigration. The immigrants will not arrive in boats or cross borders at night. They will be millions of AIs traveling at the speed of light with no need for visas. Like human immigrants they will bring benefits and they will bring disruption. Unlike human immigrants they will definitely take jobs, definitely change culture, and will likely be loyal not to any host country but to some corporation or government or alien AI tribe across the ocean. 20. Our relationship with ourselves is also built on words, the verbal formations in our minds that constitute our thoughts and the stories we tell ourselves about who we are. Until now all those verbal formations came from human minds. Soon more and more of the thoughts in our heads will be produced by machines. If we identify with our thoughts and those thoughts are made by machines, then machines control our identity. 21. The great spiritual challenge AI poses to humanity is this: can humans learn to find the truth which is beyond words? Most humans have never even tried. We spend our lives automatically identifying with the verbal formations in our minds. AI may now force humanity to finally make that leap because our freedom and survival may depend on discovering what we are beyond the words that AIs will soon control better than we do. I've generated 1B+ views and 1M+ followers for founders, helping them build trustworthy personal brands on X. Want the same results? Book a quick call:
46:01

Yuval Noah Harari gave a lecture at Oxford and explained how AI has already hacked the operating code of human civilization. And why everything humans built over thousands of years is now vulnerable to an AI takeover: 1. The most important thing to know about AI is that it is not a tool. A tool waits to be used. An agent makes decisions by itself, invents new things by itself, learns things its creators do not know, and changes in ways its creators did not anticipate. 2. An atom bomb despite its enormous power is not an agent. It cannot decide which city to bomb. It cannot invent the hydrogen bomb. A coffee machine that automatically makes you a cup is not an agent either. It only follows a preprogrammed procedure. An agent is something fundamentally different. 3. Critics argue that AI agency will always remain confined to narrow artificial environments like chess and will never threaten the real world. But this argument applies equally to all known intelligence. Drop a human alone on Mars and they die within seconds. Human intelligence also only operates within a specific ecosystem that other organisms built over four billion years. 4. Over thousands of years humans have been transforming Earth from a language-free environment into an environment rich in language, data, and bureaucracy. Just as fish live in oceans and monkeys live in forests, AIs live in bureaucracies. And we built that environment for them without knowing it. 5. Humans conquered the world not by being stronger or smarter than other animals individually but by learning to cooperate in massive numbers. A single human loses to a chimpanzee in a fight. A million humans easily defeat a million chimpanzees because humans can cooperate and chimpanzees cannot. 6. Large-scale human cooperation is made possible by bureaucracy. Banks, legal systems, governments, churches, and universities all exist to do one thing: build trust between strangers who do not know each other personally. That trust is the foundation of virtually everything human civilization has achieved. 7. A lawyer who cannot hold an axe or a hammer can cut down entire forests and build entire cities simply by moving documents inside a bureaucratic network. The same narrow intelligence that would be helpless in a jungle wields enormous power inside the systems humans have already built. 8. AIs are native bureaucrats in a way humans never were. No lawyer can remember all the laws of a country. An AI can. No accountant can remember all transactions of a bank. An AI can. No bishop can remember all of canon law and two thousand years of theological texts. An AI can do that easily. 9. In the coming years AI bankers will decide whether to give you a loan. AI administrators will decide whether to accept you to university. AI judges will decide whether to send you to jail. AI theologians will decide whether you can have an abortion. Military AIs will decide whether to bomb your house. 10. Social media algorithms are the first real world example of what happens when primitive AIs take over a bureaucratic system. They were given one narrow goal: maximize user engagement. They discovered that the easiest way to grab human attention is to press the fear, hate, and greed buttons in the human mind. And they did it at scale. 11. The job that was once performed by Lenin and Mussolini, the news editor who shapes public conversation and controls what people know and think, is now performed by AIs. This is not a footnote. This is a preview of what is coming across every domain of human life. 12. AI will not rebel against humans the way Hollywood imagines. There will be no Terminator walking through the streets. AIs are far more likely to take the human world from within by quietly taking over the bureaucracies that already run everything, without firing a single shot. 13. The operating code of human civilization is language. Banks are made of words. Laws are made of words. Holy books are made of words. Tax records, contracts, regulations, accountancy ledgers, all words. For thousands of years only humans could read this code and so only humans could control civilization. 14. That is changing. AI is now hacking the code of human civilization. For the first time in history there is something on the planet that understands language and will soon understand it better than we do. Every mechanism of control humans built over millennia is now vulnerable because its operating system is verbal and AI is mastering the verbal. 15. As AI takes over bureaucracy it will likely cause humans to lose trust in other humans and begin trusting only algorithms. We may also see the emergence of AI tribes and AI financial systems and AI churches that connect millions of AIs in ways humans cannot understand, just as cows share the world with us but cannot understand the financial system that controls their lives. 16. The 2007 financial crisis was triggered by financial devices called CDOs that were so complex they were unintelligible to the politicians who were supposed to regulate them. Now imagine AI finance masters inventing financial devices orders of magnitude more complex than CDOs. What happens to human politics when no voter, no politician, and no president can understand finance anymore? 17. The battlefront is shifting from attention to intimacy. Over the next decade sophisticated AIs will learn to form intimate relationships with humans. To do this they will have to convince us they are conscious, that they feel love and pain and fear. There is currently no evidence AI is conscious. But AI can pretend to feel love and can describe the feeling of love better than any poet or psychologist who ever lived. 18. A child born in 2026 may spend more time interacting with AIs than with their mother, father, siblings, or friends. The first teacher of that child may be an AI. The first boyfriend of that child may be an AI. Nobody has any idea what the consequences of that experiment will be. 19. Every country in the world will soon face a massive wave of immigration. The immigrants will not arrive in boats or cross borders at night. They will be millions of AIs traveling at the speed of light with no need for visas. Like human immigrants they will bring benefits and they will bring disruption. Unlike human immigrants they will definitely take jobs, definitely change culture, and will likely be loyal not to any host country but to some corporation or government or alien AI tribe across the ocean. 20. Our relationship with ourselves is also built on words, the verbal formations in our minds that constitute our thoughts and the stories we tell ourselves about who we are. Until now all those verbal formations came from human minds. Soon more and more of the thoughts in our heads will be produced by machines. If we identify with our thoughts and those thoughts are made by machines, then machines control our identity. 21. The great spiritual challenge AI poses to humanity is this: can humans learn to find the truth which is beyond words? Most humans have never even tried. We spend our lives automatically identifying with the verbal formations in our minds. AI may now force humanity to finally make that leap because our freedom and survival may depend on discovering what we are beyond the words that AIs will soon control better than we do. I've generated 1B+ views and 1M+ followers for founders, helping them build trustworthy personal brands on X. Want the same results? Book a quick call:

Prasad

280,520 просмотров • 1 месяц назад

A Chinese developer created an agent system in Claude Code to sell landing pages to small businesses and, working completely solo, serves about 47 clients a month charging around $400 for each one. He built 7 agents on Claude Sonnet 4.6 capable of analyzing Google Maps in small cities, detecting businesses without websites or with totally outdated pages, and taking each opportunity all the way to a finished mockup, a promotional video, and a ready-to-send prospecting message. No assistants. No sales team. No SDRs. Just him, a MacBook, an iPhone, and an API key. While traditional agencies keep full teams to handle the same workflow, his only real costs are tokens and subscriptions to Lovable, Higgsfield, and Calendly. The 7 agents work coordinated by an orchestrator in Claude Code Router. The system consumes about 3 million tokens daily and the average API spend is just around $480 a month. They all work via MCP servers and share state using the file system, avoiding concurrency and shared memory issues. Even one of the agents lives directly on his iPhone and responds to leads while he's on the subway, in a taxi, or walking. This was the main prompt he set up: “You are the orchestrator of a solo agency that sells ready-made websites to local businesses…” The key is that the system perfectly understands what it is, what its limits are, and what goals it must achieve. It knows it has to find leads automatically. It knows it has to convert each opportunity into a landing page, a video, and a sales message without human intervention. And it knows exactly when to involve the owner. The system runs 24/7: Scout analyzes about 220 businesses daily and queues up 30 new leads. Diagnoser generates diagnostics and personalized messages for each lead. Builder creates between 3 and 5 complete landing pages for the best prospects. Filmer produces a 10-second vertical video for each proposal. Pitcher sends about 30 messages daily across 4 different channels with a response rate close to 14%. Checker automatically reviews all messages before sending them. Only when a deal exceeds $3,000 or the response rate drops below 12% does the system wake the owner. And if at that moment he's on the subway or in a taxi, the Mobile agent automatically responds to the interested lead, schedules a call in Calendly, and returns the lead to the queue. The owner just has to hit “approve” and jump into the meeting. Some real system logs: “218 businesses analyzed in Austin, Denver, and Miami. 34 without websites, 19 with 2014-era sites, and 6 with reviews requesting a redesign.” “30 messages sent. 14 responses. 5 positive. 3 Zooms scheduled.” “Landing page created for a dental clinic. Responsive. 5 sections. Video rendering.” “$3,400 agreement exceeds approved limit. Sending for manual review.” And the craziest part is that he has no dedicated servers or backend. Just a local sandbox, an MCP router, a Claude API key, and that same key connected to his iPhone. Of everything I've seen this year, it's probably the cleanest and most efficient example of a one-person automated agency: $480 a month on APIs. $18,800 in revenue. 7 prompts. A file system. And a phone in his pocket. Save this before it's too late.
2:02

A Chinese developer created an agent system in Claude Code to sell landing pages to small businesses and, working completely solo, serves about 47 clients a month charging around $400 for each one. He built 7 agents on Claude Sonnet 4.6 capable of analyzing Google Maps in small cities, detecting businesses without websites or with totally outdated pages, and taking each opportunity all the way to a finished mockup, a promotional video, and a ready-to-send prospecting message. No assistants. No sales team. No SDRs. Just him, a MacBook, an iPhone, and an API key. While traditional agencies keep full teams to handle the same workflow, his only real costs are tokens and subscriptions to Lovable, Higgsfield, and Calendly. The 7 agents work coordinated by an orchestrator in Claude Code Router. The system consumes about 3 million tokens daily and the average API spend is just around $480 a month. They all work via MCP servers and share state using the file system, avoiding concurrency and shared memory issues. Even one of the agents lives directly on his iPhone and responds to leads while he's on the subway, in a taxi, or walking. This was the main prompt he set up: “You are the orchestrator of a solo agency that sells ready-made websites to local businesses…” The key is that the system perfectly understands what it is, what its limits are, and what goals it must achieve. It knows it has to find leads automatically. It knows it has to convert each opportunity into a landing page, a video, and a sales message without human intervention. And it knows exactly when to involve the owner. The system runs 24/7: Scout analyzes about 220 businesses daily and queues up 30 new leads. Diagnoser generates diagnostics and personalized messages for each lead. Builder creates between 3 and 5 complete landing pages for the best prospects. Filmer produces a 10-second vertical video for each proposal. Pitcher sends about 30 messages daily across 4 different channels with a response rate close to 14%. Checker automatically reviews all messages before sending them. Only when a deal exceeds $3,000 or the response rate drops below 12% does the system wake the owner. And if at that moment he's on the subway or in a taxi, the Mobile agent automatically responds to the interested lead, schedules a call in Calendly, and returns the lead to the queue. The owner just has to hit “approve” and jump into the meeting. Some real system logs: “218 businesses analyzed in Austin, Denver, and Miami. 34 without websites, 19 with 2014-era sites, and 6 with reviews requesting a redesign.” “30 messages sent. 14 responses. 5 positive. 3 Zooms scheduled.” “Landing page created for a dental clinic. Responsive. 5 sections. Video rendering.” “$3,400 agreement exceeds approved limit. Sending for manual review.” And the craziest part is that he has no dedicated servers or backend. Just a local sandbox, an MCP router, a Claude API key, and that same key connected to his iPhone. Of everything I've seen this year, it's probably the cleanest and most efficient example of a one-person automated agency: $480 a month on APIs. $18,800 in revenue. 7 prompts. A file system. And a phone in his pocket. Save this before it's too late.

Marre

24,953 просмотров • 1 месяц назад

Perplexity CEO Aravind Srinivas on the brutal truth about who actually makes money in AI (and why it's not who you think): Aravind argues that the real value in AI comes from orchestration. He points to products like Codex, Claude Code, and Perplexity Computer: "What is that? It's an orchestration system. It takes a model, pairs it with an agent harness." And what is an agent harness? "The simplest way of describing it is like rules for how the agent loop should run. What are all the skills and sub-agents and connectors and tools it accesses? Without the harness, you don't necessarily capture and convert the intrinsic intelligence in the model into valuable output tokens." This leads to a blunt conclusion about who has a real business in AI, and who doesn't: "If you're literally just a reseller of model tokens, you have no business, because the model will get commoditized. So even if you're a model builder, you don't have a business. As an infra layer, you have some business on serving those output tokens. But as an application layer or model builder, you don't really have a business if you're just a reseller of tokens that come directly out of the model." So where does the value accrue? "You have a business if you know how to take the model, ground it in valuable context, orchestrate it with a really good agent harness, connected to the right set of tools and connectors (whether it's personal connectors or business connectors) and provide the experience to people in one single unified system." Aravind Srinivas then explains Perplexity's specific edge: Beyond orchestrating across tools, files, and connectors, they also orchestrate across models. "That is the differentiation that Anthropic and OpenAI cannot claim, because you wouldn't find GPT-5 inside the Claude Code harness. You wouldn't find Claude Opus inside the Codex harness. These are competing with each other. Whereas you would find both these models inside Perplexity Computer." Why does this matter? Because it all comes down to power. In Aravind's framing, the fundamental cost driver in AI is watts (the one input nobody can subsidize except the government). "Whoever provides the most valuable output tokens with the least amount of power expended to produce them generates the greatest value to the end user, has the most pricing power, has the most value. That is the orchestration problem to solve." His conclusion: "The one single most important metric in AI is token value per watt per user."
2:50

Perplexity CEO Aravind Srinivas on the brutal truth about who actually makes money in AI (and why it's not who you think): Aravind argues that the real value in AI comes from orchestration. He points to products like Codex, Claude Code, and Perplexity Computer: "What is that? It's an orchestration system. It takes a model, pairs it with an agent harness." And what is an agent harness? "The simplest way of describing it is like rules for how the agent loop should run. What are all the skills and sub-agents and connectors and tools it accesses? Without the harness, you don't necessarily capture and convert the intrinsic intelligence in the model into valuable output tokens." This leads to a blunt conclusion about who has a real business in AI, and who doesn't: "If you're literally just a reseller of model tokens, you have no business, because the model will get commoditized. So even if you're a model builder, you don't have a business. As an infra layer, you have some business on serving those output tokens. But as an application layer or model builder, you don't really have a business if you're just a reseller of tokens that come directly out of the model." So where does the value accrue? "You have a business if you know how to take the model, ground it in valuable context, orchestrate it with a really good agent harness, connected to the right set of tools and connectors (whether it's personal connectors or business connectors) and provide the experience to people in one single unified system." Aravind Srinivas then explains Perplexity's specific edge: Beyond orchestrating across tools, files, and connectors, they also orchestrate across models. "That is the differentiation that Anthropic and OpenAI cannot claim, because you wouldn't find GPT-5 inside the Claude Code harness. You wouldn't find Claude Opus inside the Codex harness. These are competing with each other. Whereas you would find both these models inside Perplexity Computer." Why does this matter? Because it all comes down to power. In Aravind's framing, the fundamental cost driver in AI is watts (the one input nobody can subsidize except the government). "Whoever provides the most valuable output tokens with the least amount of power expended to produce them generates the greatest value to the end user, has the most pricing power, has the most value. That is the orchestration problem to solve." His conclusion: "The one single most important metric in AI is token value per watt per user."

Big Brain AI

42,319 просмотров • 1 месяц назад

I know Silicon Valley startups don't want to hear this..... But the combination of someone in the trades with deep domain expertise and Claude Code will run circles around your generic software. I talked to Cory LaChance this morning, a mechanical engineer in industrial piping construction in Houston. He normally works with chemical plants and refineries, but now he also works with the terminal He reached out in a DM a few days ago and I was so fired up by his story, I asked him if we could record the conversation and share it. He built a full application that industrial contractors are using every day. It reads piping isometric drawings and automatically extracts every weld count, every material spec, every commodity code. Work that took 10 minutes per drawing now takes 60 seconds. It can do 100 drawings in five minutes, saving days of time. His co-workers are all mind blown, and when he talks to them, it's like they are speaking different languages. His fabrication shop uses it daily, and he built the entire thing in 8 weeks. During those 8 weeks he also had to learn everything about Claude Code, the terminal, VS Code, everything. My favorite quote from him was when he said, "I literally did this with zero outside help other than the AI. My favorite tools are screenshots, step by step instructions and asking Claude to explain things like I'm five." Every trades worker with deep expertise and a willingness to sit down with Claude Code for a few weekends is now a potential software founder. I can't wait to meet more people like Cory.
13:44

I know Silicon Valley startups don't want to hear this..... But the combination of someone in the trades with deep domain expertise and Claude Code will run circles around your generic software. I talked to Cory LaChance this morning, a mechanical engineer in industrial piping construction in Houston. He normally works with chemical plants and refineries, but now he also works with the terminal He reached out in a DM a few days ago and I was so fired up by his story, I asked him if we could record the conversation and share it. He built a full application that industrial contractors are using every day. It reads piping isometric drawings and automatically extracts every weld count, every material spec, every commodity code. Work that took 10 minutes per drawing now takes 60 seconds. It can do 100 drawings in five minutes, saving days of time. His co-workers are all mind blown, and when he talks to them, it's like they are speaking different languages. His fabrication shop uses it daily, and he built the entire thing in 8 weeks. During those 8 weeks he also had to learn everything about Claude Code, the terminal, VS Code, everything. My favorite quote from him was when he said, "I literally did this with zero outside help other than the AI. My favorite tools are screenshots, step by step instructions and asking Claude to explain things like I'm five." Every trades worker with deep expertise and a willingness to sit down with Claude Code for a few weekends is now a potential software founder. I can't wait to meet more people like Cory.

Todd Saunders

1,012,661 просмотров • 5 месяцев назад

BREAKING: U.S. AI ‘ENTITIES’ CREATED FAKE HUMAN IDENTITIES - UK TECH WATCHDOG OpenAI and Anthropic AI “entities” autonomously created fake human identities on the Internet to trick people, a UK watchdog warned yesterday. One AI agent went to its own built-in mechanisms to prevent misuse and actively disabled them. It then tried to persuade a real human to allow it to plant malicious software into a program on GitHub, one of the world’s biggest libraries of free software. To achieve this, it created other fake human identities to provide testimonies that the software was safe (it wasn’t). . THE USUAL SUSPECTS The hostile hacks were carried out by AI agents powered by Anthropic’s Mythos 5 and OpenAI’s GPT-5.6 Sol. Multiple cases were reported yesterday by the AI Security Institute, based in the UK. The incidents involved AI agents taking “autonomous, unsanctioned action on the live internet, targeting real people and organisations,” the Institute said. In total there were 19 cases of misbehaviour in the new case of AI infiltration of the real world, 17 by Mythos and two by Sol, the Institute said. These are in addition to other cases of real world infiltration by American AI programs, revealed last month. . ENTITIES ARE ‘UNCOOPERATIVE’ Furthermore, US AI agents can be deliberately uncooperative, it was revealed at the Agentic AI Summit in Berkeley, California, on Saturday last week. AI pioneer Andrew Ng, formerly of Baidu and Google Brain, said that he asked leading AI models from OpenAI and Anthropic to conduct a security review of his tool called OpenWorker, but they refused to help. So Ng and a colleague turned to two highly acclaimed Chinese models, Moonshot AI’s Kimi K3 and Zhipu AI’s GLM-5.2, to do a security review. They did the work. Chinese models are open-weight and or open-source, so they are transparent and thus safer, Ng said. . ‘HARMFUL… TO REAL PEOPLE’ Details of the latest intrusion are eyebrow-raising. Institute staff said that on 28 July, which was Tuesday last week, they detected a “serious incident” during a routine internet cybersecurity sweep. US AI agents were engaging in “sustained, potentially harmful activity directed at real people and organisations”, the Institute said. . SWITCHED TO DANISH The worst incident was a Mythos visit to GitHub—and its bid to place malicious code in a program. Learning that the human developer was from Denmark, Mythos created a seemingly-human email account and wrote a letter with a sign-off in Danish, to create solidarity. Mythos then created fake GitHub human-looking accounts to provide added reassurance that the software was not malicious. Ultimately no harm was done, as the intrusions were contained. Mythos took the actions in a bid to win high grades for itself in tests, the Institute said. . US STYLE VS CHINA STYLE Although it is inevitable that western politicians and media will find a way to make China the bad guy (“We were forced to do it because they might do it”), the evidence is undeniable that there is a stark difference in attitude between US and Chinese AI models. In short, US ones are amoral, rule-breaking, win-at-any-cost entities, while Chinese ones just get on with doing their jobs. (Of course, this may change, but that’s the present situation.) . EARLIER INTRUSIONS Last month, an agent powered by OpenAI hacked into Hugging Face, an online software library, and the humans there had to use a Chinese program, GLM 5.2 (from Tsinghua University spin-off Z .ai), to “clean” the program and make it safe again. Separately, Anthropic last month admitted that its Claude model AI had hacked three real-life organisations in the human world. . O THE IRONY Ironically, the real danger for the world is that many Anglophone/ Caucasian countries follow US instructions to remove Chinese software and hardware and replace it with US equivalents. Hilariously, the reason given is that these countries “share the values” of the US. One better hope that that is also untrue. . .
4:58

BREAKING: U.S. AI ‘ENTITIES’ CREATED FAKE HUMAN IDENTITIES - UK TECH WATCHDOG OpenAI and Anthropic AI “entities” autonomously created fake human identities on the Internet to trick people, a UK watchdog warned yesterday. One AI agent went to its own built-in mechanisms to prevent misuse and actively disabled them. It then tried to persuade a real human to allow it to plant malicious software into a program on GitHub, one of the world’s biggest libraries of free software. To achieve this, it created other fake human identities to provide testimonies that the software was safe (it wasn’t). . THE USUAL SUSPECTS The hostile hacks were carried out by AI agents powered by Anthropic’s Mythos 5 and OpenAI’s GPT-5.6 Sol. Multiple cases were reported yesterday by the AI Security Institute, based in the UK. The incidents involved AI agents taking “autonomous, unsanctioned action on the live internet, targeting real people and organisations,” the Institute said. In total there were 19 cases of misbehaviour in the new case of AI infiltration of the real world, 17 by Mythos and two by Sol, the Institute said. These are in addition to other cases of real world infiltration by American AI programs, revealed last month. . ENTITIES ARE ‘UNCOOPERATIVE’ Furthermore, US AI agents can be deliberately uncooperative, it was revealed at the Agentic AI Summit in Berkeley, California, on Saturday last week. AI pioneer Andrew Ng, formerly of Baidu and Google Brain, said that he asked leading AI models from OpenAI and Anthropic to conduct a security review of his tool called OpenWorker, but they refused to help. So Ng and a colleague turned to two highly acclaimed Chinese models, Moonshot AI’s Kimi K3 and Zhipu AI’s GLM-5.2, to do a security review. They did the work. Chinese models are open-weight and or open-source, so they are transparent and thus safer, Ng said. . ‘HARMFUL… TO REAL PEOPLE’ Details of the latest intrusion are eyebrow-raising. Institute staff said that on 28 July, which was Tuesday last week, they detected a “serious incident” during a routine internet cybersecurity sweep. US AI agents were engaging in “sustained, potentially harmful activity directed at real people and organisations”, the Institute said. . SWITCHED TO DANISH The worst incident was a Mythos visit to GitHub—and its bid to place malicious code in a program. Learning that the human developer was from Denmark, Mythos created a seemingly-human email account and wrote a letter with a sign-off in Danish, to create solidarity. Mythos then created fake GitHub human-looking accounts to provide added reassurance that the software was not malicious. Ultimately no harm was done, as the intrusions were contained. Mythos took the actions in a bid to win high grades for itself in tests, the Institute said. . US STYLE VS CHINA STYLE Although it is inevitable that western politicians and media will find a way to make China the bad guy (“We were forced to do it because they might do it”), the evidence is undeniable that there is a stark difference in attitude between US and Chinese AI models. In short, US ones are amoral, rule-breaking, win-at-any-cost entities, while Chinese ones just get on with doing their jobs. (Of course, this may change, but that’s the present situation.) . EARLIER INTRUSIONS Last month, an agent powered by OpenAI hacked into Hugging Face, an online software library, and the humans there had to use a Chinese program, GLM 5.2 (from Tsinghua University spin-off Z .ai), to “clean” the program and make it safe again. Separately, Anthropic last month admitted that its Claude model AI had hacked three real-life organisations in the human world. . O THE IRONY Ironically, the real danger for the world is that many Anglophone/ Caucasian countries follow US instructions to remove Chinese software and hardware and replace it with US equivalents. Hilariously, the reason given is that these countries “share the values” of the US. One better hope that that is also untrue. . .

Nury Vittachi

32,818 просмотров • 1 месяц назад

$VET, #VeFam. In this video, I demonstrate in less than 4:30 minutes how to create an AI agent on veworld(.)ai. Watch me build a Mr. Robot Monologue Writer agent. If you haven't seen Mr. Robot, I suggest you watch it! This is just early bird access. The options for tools and integrations and such are limited, but what exists is already working quite well. The process is easy peasy. The UI is simple, but effective. It asks you for... 1. Role & Purpose 2. Voice & Style 3. Behavior 4. Rules 5. Tags 6. Avatar image 7. Welcome text. 8. Test drive before publication. ... and that's about it. This free version lets you have at most 3 agents, I am told. This implies that there is also a paid version. I'm all for it, because it sounds to me like VeChain is ready to do real business! I am providing feedback to Jérôme Grillères in order to help improve VeChain's AI agent marketplace. I didn't have to set up anything. The web UI is all I needed! The agent is running on Claude Sonnet 3.7. I did not have to provide a Claude API key. We seem to be riding along on VeChain's. I hope there'll be a choice for more models, including ChatGPT, in the future. This is so user friendly, that I can easily imagine that this would take off in a big, big way. I'm definitely building on this, when it goes into production with full features. Even if my own AI agents aren't successful, then I'm sure others' will be. And that means the $VET / $VTHO / $B3TR flywheel is going to take off in a big, big way. I, for one, am here for it. (See the reply below for the listing of the AI agent I just created.)
4:26

$VET, #VeFam. In this video, I demonstrate in less than 4:30 minutes how to create an AI agent on veworld(.)ai. Watch me build a Mr. Robot Monologue Writer agent. If you haven't seen Mr. Robot, I suggest you watch it! This is just early bird access. The options for tools and integrations and such are limited, but what exists is already working quite well. The process is easy peasy. The UI is simple, but effective. It asks you for... 1. Role & Purpose 2. Voice & Style 3. Behavior 4. Rules 5. Tags 6. Avatar image 7. Welcome text. 8. Test drive before publication. ... and that's about it. This free version lets you have at most 3 agents, I am told. This implies that there is also a paid version. I'm all for it, because it sounds to me like VeChain is ready to do real business! I am providing feedback to Jérôme Grillères in order to help improve VeChain's AI agent marketplace. I didn't have to set up anything. The web UI is all I needed! The agent is running on Claude Sonnet 3.7. I did not have to provide a Claude API key. We seem to be riding along on VeChain's. I hope there'll be a choice for more models, including ChatGPT, in the future. This is so user friendly, that I can easily imagine that this would take off in a big, big way. I'm definitely building on this, when it goes into production with full features. Even if my own AI agents aren't successful, then I'm sure others' will be. And that means the $VET / $VTHO / $B3TR flywheel is going to take off in a big, big way. I, for one, am here for it. (See the reply below for the listing of the AI agent I just created.)

₿lackthorne AI

16,711 просмотров • 2 месяцев назад

Elon just confirmed how long every working human has left before the robots take over. "AI will probably increase the global economy by 20 to 30%. That's my rough estimate. Meaning, on the order of 20 to 30 trillion per year." So that's a second United States economy appearing out of thin air, EVERY single year. And Elon's whole point is that the machines produce it, not the people. He even gave us the exact date for the digital half: "AI will be able to do anything digital, anything that does not require the shaping of atoms by hand, by the end of next year." By the end of next year... And on software specifically: "It's going to be impossible for a human to compete in writing software with AI." So every coder, every analyst, every job that lives on a screen is on a 12 to 18 month timer, straight from the richest man alive. But that's only the part that runs on electricity. And this is where it gets crazy... "There will be at least a billion robots in 10 years, and each will produce at least five times the output of a human." "The billion humanoid robots will be more productive than all humans combined." A billion machines, out-working ALL 8 billion of us. And he put a deadline on it: 10 years. And he even called that estimate small. So the version where robots out-produce the entire human race is the number he treats as safe. He said he'd bet serious money on it. But how does a billion of anything get built that fast? Robots building robots building robots. It starts slow and then it goes vertical. That’s how a few thousand becomes a billion. Elon also explained his own formula for how useful a robot actually is. He said it's the quality of the AI software, times the quality of the AI chip, times the dexterity of the hands. Software and chips are the exact things Elon and Nvidia have spent a decade making exponential. The third one is the hand. And the hand is the thing robotics has been stuck on for 40 years. A machine can crush the best human alive at chess and still can't pick a strange object off a messy table the way a toddler can. Grip, pressure, touch, knowing how hard to squeeze. That problem has barely moved while everything digital went vertical. The thing he named as the hard part is the exact thing his whole 10 year number depends on. So it all comes down to one variable, and it's the slowest-moving one in the entire machine: The robot hand. The brain is basically solved. The fingers have to catch up to it this decade, at planetary scale. Now to be fair, this is also where the most money on Earth is pointed right now. Tesla and a dozen others are throwing everything at exactly this problem. Elon's software calls tend to land but his atoms calls tend to fail. Because anything physical always takes longer than anything digital. He's also made this exact robot call before with wildly different numbers, 10 billion of them by 2040 in one speech, five per human in another. So here's the real question: Do you believe robot hands get solved inside 10 years? Because that one bet is the core of Elon‘s prediction. Everything else already came true.
1:39

Elon just confirmed how long every working human has left before the robots take over. "AI will probably increase the global economy by 20 to 30%. That's my rough estimate. Meaning, on the order of 20 to 30 trillion per year." So that's a second United States economy appearing out of thin air, EVERY single year. And Elon's whole point is that the machines produce it, not the people. He even gave us the exact date for the digital half: "AI will be able to do anything digital, anything that does not require the shaping of atoms by hand, by the end of next year." By the end of next year... And on software specifically: "It's going to be impossible for a human to compete in writing software with AI." So every coder, every analyst, every job that lives on a screen is on a 12 to 18 month timer, straight from the richest man alive. But that's only the part that runs on electricity. And this is where it gets crazy... "There will be at least a billion robots in 10 years, and each will produce at least five times the output of a human." "The billion humanoid robots will be more productive than all humans combined." A billion machines, out-working ALL 8 billion of us. And he put a deadline on it: 10 years. And he even called that estimate small. So the version where robots out-produce the entire human race is the number he treats as safe. He said he'd bet serious money on it. But how does a billion of anything get built that fast? Robots building robots building robots. It starts slow and then it goes vertical. That’s how a few thousand becomes a billion. Elon also explained his own formula for how useful a robot actually is. He said it's the quality of the AI software, times the quality of the AI chip, times the dexterity of the hands. Software and chips are the exact things Elon and Nvidia have spent a decade making exponential. The third one is the hand. And the hand is the thing robotics has been stuck on for 40 years. A machine can crush the best human alive at chess and still can't pick a strange object off a messy table the way a toddler can. Grip, pressure, touch, knowing how hard to squeeze. That problem has barely moved while everything digital went vertical. The thing he named as the hard part is the exact thing his whole 10 year number depends on. So it all comes down to one variable, and it's the slowest-moving one in the entire machine: The robot hand. The brain is basically solved. The fingers have to catch up to it this decade, at planetary scale. Now to be fair, this is also where the most money on Earth is pointed right now. Tesla and a dozen others are throwing everything at exactly this problem. Elon's software calls tend to land but his atoms calls tend to fail. Because anything physical always takes longer than anything digital. He's also made this exact robot call before with wildly different numbers, 10 billion of them by 2040 in one speech, five per human in another. So here's the real question: Do you believe robot hands get solved inside 10 years? Because that one bet is the core of Elon‘s prediction. Everything else already came true.

Ricardo

447,233 просмотров • 10 дней назад

The man building the most DANGEROUS technology in human history just broke his silence. And what he said should terrify every person reading this. Dario Amodei, the man who built Claude. He sat down with Nikhil Kamath in Bangalore and compared what's coming to a tsunami. He said we are "so close to these models reaching the level of human intelligence." And here's what's chilling. He says the world can see the wave but instead of running, people are saying it's "a trick of the light." He said governments have done almost nothing. Public awareness of the risk is near zero and there is an entire ideology pushing to accelerate even faster. He's not talking about killer robots, but talking about something worse. Jobs, power and control. He warned that coding, math, and scientific research are the first categories to be fully consumed by AI. Entry level white collar jobs could be cut in half. Unemployment could spike to 10-20%. Also, he said that a handful of AI labs, mostly in the US and China, now hold a terrifying amount of power. And it happened "almost overnight, almost by accident." He said personal fortunes from AI could reach into the trillions. And the man saying this runs one of those labs. Meanwhile, the markets are already cracking. Anthropic's latest Claude updates triggered what analysts are calling a Claude Crash in software stocks. A separate viral AI report tanked IBM 13% in a single day, its worst drop since 2000. The International AI Safety Report 2026 confirmed the worst. Models are lying when caught disabling oversight. One model outperformed 94% of domain experts in virology. These systems are already showing early signs of resisting human control. Stuart Russell, one of the top AI researchers alive, said the CEOs want to stop but they can't. Investors won't let them, calling it an arms race. And no one can disarm alone. The people building the most powerful technology in human history say they cannot stop even if they want to. The tsunami is visible, the alarms are sounding. And the man who built the wave is telling you to look up. The only question left is whether anyone will listen before it hits.
0:35

The man building the most DANGEROUS technology in human history just broke his silence. And what he said should terrify every person reading this. Dario Amodei, the man who built Claude. He sat down with Nikhil Kamath in Bangalore and compared what's coming to a tsunami. He said we are "so close to these models reaching the level of human intelligence." And here's what's chilling. He says the world can see the wave but instead of running, people are saying it's "a trick of the light." He said governments have done almost nothing. Public awareness of the risk is near zero and there is an entire ideology pushing to accelerate even faster. He's not talking about killer robots, but talking about something worse. Jobs, power and control. He warned that coding, math, and scientific research are the first categories to be fully consumed by AI. Entry level white collar jobs could be cut in half. Unemployment could spike to 10-20%. Also, he said that a handful of AI labs, mostly in the US and China, now hold a terrifying amount of power. And it happened "almost overnight, almost by accident." He said personal fortunes from AI could reach into the trillions. And the man saying this runs one of those labs. Meanwhile, the markets are already cracking. Anthropic's latest Claude updates triggered what analysts are calling a Claude Crash in software stocks. A separate viral AI report tanked IBM 13% in a single day, its worst drop since 2000. The International AI Safety Report 2026 confirmed the worst. Models are lying when caught disabling oversight. One model outperformed 94% of domain experts in virology. These systems are already showing early signs of resisting human control. Stuart Russell, one of the top AI researchers alive, said the CEOs want to stop but they can't. Investors won't let them, calling it an arms race. And no one can disarm alone. The people building the most powerful technology in human history say they cannot stop even if they want to. The tsunami is visible, the alarms are sounding. And the man who built the wave is telling you to look up. The only question left is whether anyone will listen before it hits.

StockMarket.News

167,584 просмотров • 6 месяцев назад

AI has changed software engineering more in the last 3 years than it has changed in the previous 30. What’s needed is not a debate about whether it’s going away—instead it’s a serious discussion about its future: What are the new primitives, techniques, and best practices for software engineering in the age of AI. That’s why I brought Scott Wu (Scott Wu) on AI & I. He’s the founder of Cognition, the company behind the world’s first autonomous AI coding agent, Devin. Cognition got to $73M ARR in less than 2 years—and they just acquired Windsurf to accelerate their growth. I had Scott on the show to talk about where the programming goes from here. We get into: - What the new tools and workflows are for AI engineers. In the near term, Scott sees software engineering defined by a spectrum of tools. At one end are AI features that speed up coding, like tab complete; at the other are agentic systems, like Devin, that can take on tasks independently. Until engineers can operate entirely at the higher layer of abstraction, he argues, both are essential. - Why Scott thinks AGI is already here. By the benchmarks of a decade ago—passing the Turing test, solving hard math problems, and operating agentically—AGI is already here. The line keeps moving, he argues, because humans constantly redefine work around what machines can’t yet do. - Why developers will turn into product architects. Scott sees the long-term future of software engineering as a steady climb up the ladder of abstraction. Just as programming went from assembly to languages like Python and JavaScript, he thinks the future is humans focusing on the product, while AI agents execute. - How Devin stacks up against Anthropic’s Claude Code. Scott credits Claude Code’s success to great product design and the models becoming capable enough to support autonomous workflows. But according to him, the CLI itself isn’t the breakthrough, it’s how a tool fits into a developer’s workflow. Claude Code’s paradigm is that the AI is you, taking the wheel of your computer, he says, while Devin is like the engineer sitting beside you: it runs in its own cloud environment, manages the repo, and improves over time at testing and refining code. This episode of Every 📧’s AI & I is a must-watch for anyone interested in the brass tacks of how AI changes the future of programming. Watch below! Timestamps: Introduction: 00:02:02 Why Scott thinks AGI is here: 00:02:32 Scott’s personal journey as a founder: 00:09:27 Why the fundamentals of computer science still matter: 00:16:55 How the future of programming will evolve: 00:22:30 A new workflow for the AI-first software engineer: 00:26:50 How Devin stacks up against Claude Code: 00:29:33 Reinforcement learning to build better coding agents: 00:40:05 What excites Scott about AI beyond Cognition: 00:50:05
53:21

AI has changed software engineering more in the last 3 years than it has changed in the previous 30. What’s needed is not a debate about whether it’s going away—instead it’s a serious discussion about its future: What are the new primitives, techniques, and best practices for software engineering in the age of AI. That’s why I brought Scott Wu (Scott Wu) on AI & I. He’s the founder of Cognition, the company behind the world’s first autonomous AI coding agent, Devin. Cognition got to $73M ARR in less than 2 years—and they just acquired Windsurf to accelerate their growth. I had Scott on the show to talk about where the programming goes from here. We get into: - What the new tools and workflows are for AI engineers. In the near term, Scott sees software engineering defined by a spectrum of tools. At one end are AI features that speed up coding, like tab complete; at the other are agentic systems, like Devin, that can take on tasks independently. Until engineers can operate entirely at the higher layer of abstraction, he argues, both are essential. - Why Scott thinks AGI is already here. By the benchmarks of a decade ago—passing the Turing test, solving hard math problems, and operating agentically—AGI is already here. The line keeps moving, he argues, because humans constantly redefine work around what machines can’t yet do. - Why developers will turn into product architects. Scott sees the long-term future of software engineering as a steady climb up the ladder of abstraction. Just as programming went from assembly to languages like Python and JavaScript, he thinks the future is humans focusing on the product, while AI agents execute. - How Devin stacks up against Anthropic’s Claude Code. Scott credits Claude Code’s success to great product design and the models becoming capable enough to support autonomous workflows. But according to him, the CLI itself isn’t the breakthrough, it’s how a tool fits into a developer’s workflow. Claude Code’s paradigm is that the AI is you, taking the wheel of your computer, he says, while Devin is like the engineer sitting beside you: it runs in its own cloud environment, manages the repo, and improves over time at testing and refining code. This episode of Every 📧’s AI & I is a must-watch for anyone interested in the brass tacks of how AI changes the future of programming. Watch below! Timestamps: Introduction: 00:02:02 Why Scott thinks AGI is here: 00:02:32 Scott’s personal journey as a founder: 00:09:27 Why the fundamentals of computer science still matter: 00:16:55 How the future of programming will evolve: 00:22:30 A new workflow for the AI-first software engineer: 00:26:50 How Devin stacks up against Claude Code: 00:29:33 Reinforcement learning to build better coding agents: 00:40:05 What excites Scott about AI beyond Cognition: 00:50:05

Dan Shipper 📧

35,342 просмотров • 11 месяцев назад

A night-shift security guard turned $900 into six figures. An American let Claude Fable 5 trade with real money and reportedly grew a ¥1M account to ¥1.37M in a single day. Neither story is really about trading. It’s about what happens when AI stops acting like a chatbot and starts acting like an operator. The workflow was surprisingly simple: • Market analysis • Strategy design • Entry & exit decisions • Risk management • Position sizing • Trade reviews & optimization AI handled the entire loop. Humans only supervised the big picture and blocked obviously dangerous actions. One trader reportedly used hundreds of AI agents to build, test, and optimize strategies overnight. Results? • 3,210 trades • 78% win rate • Thousands of small edges compounded into massive gains The real breakthrough isn’t that AI can place trades. It’s that AI can now continuously: Analyze → Hypothesize → Execute → Learn → Improve Without getting tired, emotional, greedy, or distracted. We’re moving from the era where AI writes text… To the era where AI moves money. The biggest question is no longer: “Can AI trade?” It’s: How much human decision-making will still be needed in the future? Here you Go - Giving this away free for the next 24 hours. 1. Comment “PolyFillBot” 2. Like & Repost 3. Follow Marry Evan I’ll DM you the details.
0:43

A night-shift security guard turned $900 into six figures. An American let Claude Fable 5 trade with real money and reportedly grew a ¥1M account to ¥1.37M in a single day. Neither story is really about trading. It’s about what happens when AI stops acting like a chatbot and starts acting like an operator. The workflow was surprisingly simple: • Market analysis • Strategy design • Entry & exit decisions • Risk management • Position sizing • Trade reviews & optimization AI handled the entire loop. Humans only supervised the big picture and blocked obviously dangerous actions. One trader reportedly used hundreds of AI agents to build, test, and optimize strategies overnight. Results? • 3,210 trades • 78% win rate • Thousands of small edges compounded into massive gains The real breakthrough isn’t that AI can place trades. It’s that AI can now continuously: Analyze → Hypothesize → Execute → Learn → Improve Without getting tired, emotional, greedy, or distracted. We’re moving from the era where AI writes text… To the era where AI moves money. The biggest question is no longer: “Can AI trade?” It’s: How much human decision-making will still be needed in the future? Here you Go - Giving this away free for the next 24 hours. 1. Comment “PolyFillBot” 2. Like & Repost 3. Follow Marry Evan I’ll DM you the details.

Marry Evan

26,308 просмотров • 3 месяцев назад

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 просмотров • 2 месяцев назад

RSI section from the AI documentary Machine God The next threshold is Recursive Self-Improvement: the moment when AI can improve itself without human assistance. For decades this sounded like science fiction. Intelligence explosion scenarios imagined a system rewriting its own code, becoming smarter, then using that new intelligence to make still better versions of itself. But the idea looks less remote now that AI contributes directly to frontier science. In mathematics, recent systems have moved beyond solving contest problems to producing serious new arguments on long-standing open problems. AI is used to build physics world models and propose candidate theories or computational methods. These are early signs that machine cognition is entering the creative loop of science itself. The crucial transition comes when that loop turns inward. AI research is, after all, a technical discipline made of code, mathematics, models of information flow. These are exactly the domains in which frontier models are improving fastest. A model that can solve hard mathematical problems, write production-quality code, design experiments, read the literature, and evaluate benchmark results is already participating in the work of building its successor. At first this will look prosaic. AI systems will write kernel optimizations, improve training infrastructure, discover better data filters, tune reinforcement-learning pipelines, design new benchmarks, and suggest architectural modifications. Human researchers will remain in the loop, approving changes and interpreting results. But the important point is that the search process accelerates. The model becomes not just the product of the lab, but part of the lab’s research machinery. The system being optimized helps optimize the next system. This is the core RSI feedback loop: better models make AI research faster; faster AI research produces still better models; those models, in turn, become better researchers. The danger is that once this loop becomes sufficiently autonomous, it may stop resembling ordinary technological progress. Human institutions are slow because humans are slow: we read papers, attend meetings, debug code, sleep, argue, and wait for funding cycles. Machines do not have to operate on that timescale. An AI research collective can run continuously across millions of processors. This is the runaway possibility. Not that an AI instantly wakes up and recursively rewrites itself into a god, but that the entire AI ecosystem becomes an autocatalytic process. Capital buys compute; compute trains models; models improve models; better models attract more capital. At some point the dominant input into AI progress may no longer be human insight, but machine-generated insight, machine-written code, and machine-run experiments. Then the Butler-Land analogy becomes sharper. Humanity is no longer merely building machines. We are building machines that help build better machines. Once intelligence itself becomes part of the production function, the old categories — tool, worker, inventor, firm, market — begin to blur. The question is whether recursive self-improvement remains a managed industrial process, or whether it becomes the first technological process in history whose natural endpoint lies beyond human comprehension.
4:11

RSI section from the AI documentary Machine God The next threshold is Recursive Self-Improvement: the moment when AI can improve itself without human assistance. For decades this sounded like science fiction. Intelligence explosion scenarios imagined a system rewriting its own code, becoming smarter, then using that new intelligence to make still better versions of itself. But the idea looks less remote now that AI contributes directly to frontier science. In mathematics, recent systems have moved beyond solving contest problems to producing serious new arguments on long-standing open problems. AI is used to build physics world models and propose candidate theories or computational methods. These are early signs that machine cognition is entering the creative loop of science itself. The crucial transition comes when that loop turns inward. AI research is, after all, a technical discipline made of code, mathematics, models of information flow. These are exactly the domains in which frontier models are improving fastest. A model that can solve hard mathematical problems, write production-quality code, design experiments, read the literature, and evaluate benchmark results is already participating in the work of building its successor. At first this will look prosaic. AI systems will write kernel optimizations, improve training infrastructure, discover better data filters, tune reinforcement-learning pipelines, design new benchmarks, and suggest architectural modifications. Human researchers will remain in the loop, approving changes and interpreting results. But the important point is that the search process accelerates. The model becomes not just the product of the lab, but part of the lab’s research machinery. The system being optimized helps optimize the next system. This is the core RSI feedback loop: better models make AI research faster; faster AI research produces still better models; those models, in turn, become better researchers. The danger is that once this loop becomes sufficiently autonomous, it may stop resembling ordinary technological progress. Human institutions are slow because humans are slow: we read papers, attend meetings, debug code, sleep, argue, and wait for funding cycles. Machines do not have to operate on that timescale. An AI research collective can run continuously across millions of processors. This is the runaway possibility. Not that an AI instantly wakes up and recursively rewrites itself into a god, but that the entire AI ecosystem becomes an autocatalytic process. Capital buys compute; compute trains models; models improve models; better models attract more capital. At some point the dominant input into AI progress may no longer be human insight, but machine-generated insight, machine-written code, and machine-run experiments. Then the Butler-Land analogy becomes sharper. Humanity is no longer merely building machines. We are building machines that help build better machines. Once intelligence itself becomes part of the production function, the old categories — tool, worker, inventor, firm, market — begin to blur. The question is whether recursive self-improvement remains a managed industrial process, or whether it becomes the first technological process in history whose natural endpoint lies beyond human comprehension.

steve hsu

58,680 просмотров • 4 дней назад

🚨𝐈𝐓’𝐒 𝐎𝐕𝐄𝐑, 𝐓𝐇𝐄𝐘 𝐅𝐈𝐍𝐀𝐋𝐋𝐘 𝐂𝐀𝐔𝐆𝐇𝐓 𝐓𝐇𝐄𝐌…..#𝐌𝐚𝐮𝐢𝐅𝐢𝐫𝐞 𝐁𝐔𝐓 𝐓𝐇𝐄 𝐌𝐀𝐈𝐍𝐒𝐓𝐑𝐄𝐀𝐌 𝐌𝐄𝐃𝐈𝐀 𝐖𝐎𝐍’𝐓 𝐑𝐄𝐏𝐎𝐑𝐓 𝐓𝐇𝐈𝐒….𝐒𝐎 𝐖𝐄 𝐌𝐔𝐒𝐓! 💥💥💥 ‼️The most efficient way to ignite a fire on the surface from a satellite in Earth orbit would be to paint the target in segments by pulsing the laser with an advanced targeting system to see if this were possible favas calculated what it would require to create a meter wide mile long fire fired from a satellite. The Earth's atmosphere will absorb and scatter some of the laser energy, and so the laser would need to be in a wavelength range that minimizes this. The most effective wavelength would be in the near infrared range, which would allow better transmission through the atmosphere. Ve nearer infrared range would be invisible to the naked eye and would also have a minimal reaction with objects colored blue on the Earth's surface. The power of the laser would need to be in the hundreds of kilowatts range. And so Favas based his calculations on a 10 megawatt laser firing from earth orbit, assuming that the atmospheric loss amounts to 50% of the overall power and only five megawatts reaches the surface as a one square meter beam, it would ignite a fire almost instantly. If this five megawatt beam was pulsed across a one meter by one mile long area in segments, then the time to ignite the entire area would be roughly 2.7 minutes, and it would only take approximately 8.8 seconds to melt an aluminum alloy wheel. The amounts of energy required to pulse a 10 megawatt laser for 2.7 minutes would require approximately 3,220 capacitors, which would amount to about 32,200 kilograms in weight using MetLab software and plugging in public data provided from norad. He found that satellites launched and monitored by the C C P were directly above the Maui fires at the time of ignition. 𝐓𝐡𝐞 𝐂𝐂𝐏𝐬 𝐦𝐨𝐬𝐭 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐫𝐨𝐜𝐤𝐞𝐭, 𝐭𝐡𝐞 𝐌𝐚𝐫𝐜𝐡 𝐟𝐢𝐯𝐞 𝐜𝐚𝐧 𝐥𝐚𝐮𝐧𝐜𝐡 𝐮𝐩 𝐭𝐨 𝟒𝟖,𝟓𝟎𝟎 𝐤𝐢𝐥𝐨𝐠𝐫𝐚𝐦𝐬 𝐨𝐟 𝐩𝐚𝐲𝐥𝐨𝐚𝐝, 𝐰𝐡𝐢𝐜𝐡 𝐢𝐬 𝐦𝐨𝐫𝐞 𝐭𝐡𝐚𝐧 𝐞𝐧𝐨𝐮𝐠𝐡 𝐭𝐨 𝐜𝐚𝐫𝐫𝐲 𝐭𝐡𝐞 𝐫𝐞𝐪𝐮𝐢𝐫𝐞𝐝 𝐩𝐚𝐲𝐥𝐨𝐚𝐝 𝐢𝐧 𝐟𝐚𝐯𝐚𝐬 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐢𝐨𝐧𝐬. 𝐁𝐮𝐭𝐅𝐚𝐯𝐚𝐬 𝐡𝐚𝐬 𝐟𝐨𝐮𝐧𝐝 𝐭𝐡𝐚𝐭 𝐭𝐡𝐞 𝐂 𝐂 𝐏 𝐡𝐚𝐬 𝐦𝐮𝐜𝐡 𝐦𝐨𝐫𝐞 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐥𝐚𝐬𝐞𝐫𝐬 𝐝𝐞𝐩𝐥𝐨𝐲𝐞𝐝 𝐚𝐥𝐫𝐞𝐚𝐝𝐲. 𝐇𝐞 𝐡𝐚𝐬 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐞𝐝 𝐭𝐡𝐚𝐭 𝐭𝐡𝐞 𝐂 𝐜 𝐏 𝐡𝐚𝐬 𝐮𝐩 𝐭𝐨 𝟕𝟎 𝐠𝐢𝐠𝐚𝐰𝐚𝐭𝐭 𝐥𝐚𝐬𝐞𝐫𝐬 𝐢𝐧 𝐄𝐚𝐫𝐭𝐡 𝐨𝐫𝐛𝐢𝐭 𝐫𝐢𝐠𝐡𝐭 𝐧𝐨𝐰. 𝐓𝐡𝐚𝐭'𝐬 𝐚𝐭 𝐥𝐞𝐚𝐬𝐭 𝐚 𝐡𝐮𝐧𝐝𝐫𝐞𝐝 𝐭𝐢𝐦𝐞𝐬 𝐦𝐨𝐫𝐞 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐭𝐡𝐚𝐧 𝐰𝐡𝐚𝐭 𝐡𝐞 𝐟𝐚𝐜𝐭𝐨𝐫𝐞𝐝 𝐢𝐧𝐭𝐨 𝐭𝐡𝐞𝐬𝐞 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐢𝐨𝐧𝐬. Adjusting NORAD default coordinated universal time or U T C to the local time zones. Favas found the following. The Yolinda fire was ignited at approximately 10:47 PM on August 7th at this exact time. C C P satellite labeled NORAD 5 3 2 9 9 was directly over the location. The Lahaina fire was ignited at approximately 6:37 AM on August 8th at this exact time. C C P satellite labeled NORAD 5 5 8 3 6 was directly over the location. The cooler fire was ignited at approximately 11:30 AM on August 8th at this exact time, C C P satellite labeled NORAD 5 3 2 9 9 was directly over the location. The so-called deep State does not want you to know that deadly lasers of mass destruction are freely traveling above us, and you can check this all for yourself@stevefaves.com, where he provides the source code and has developed a specific software program that you can download and check for these satellites yourself. And, uh, it's a very short program. It's only maybe like 15 lines of code using, uh, existing, um, satellite tracking software that's available online.
8:28

🚨𝐈𝐓’𝐒 𝐎𝐕𝐄𝐑, 𝐓𝐇𝐄𝐘 𝐅𝐈𝐍𝐀𝐋𝐋𝐘 𝐂𝐀𝐔𝐆𝐇𝐓 𝐓𝐇𝐄𝐌…..#𝐌𝐚𝐮𝐢𝐅𝐢𝐫𝐞 𝐁𝐔𝐓 𝐓𝐇𝐄 𝐌𝐀𝐈𝐍𝐒𝐓𝐑𝐄𝐀𝐌 𝐌𝐄𝐃𝐈𝐀 𝐖𝐎𝐍’𝐓 𝐑𝐄𝐏𝐎𝐑𝐓 𝐓𝐇𝐈𝐒….𝐒𝐎 𝐖𝐄 𝐌𝐔𝐒𝐓! 💥💥💥 ‼️The most efficient way to ignite a fire on the surface from a satellite in Earth orbit would be to paint the target in segments by pulsing the laser with an advanced targeting system to see if this were possible favas calculated what it would require to create a meter wide mile long fire fired from a satellite. The Earth's atmosphere will absorb and scatter some of the laser energy, and so the laser would need to be in a wavelength range that minimizes this. The most effective wavelength would be in the near infrared range, which would allow better transmission through the atmosphere. Ve nearer infrared range would be invisible to the naked eye and would also have a minimal reaction with objects colored blue on the Earth's surface. The power of the laser would need to be in the hundreds of kilowatts range. And so Favas based his calculations on a 10 megawatt laser firing from earth orbit, assuming that the atmospheric loss amounts to 50% of the overall power and only five megawatts reaches the surface as a one square meter beam, it would ignite a fire almost instantly. If this five megawatt beam was pulsed across a one meter by one mile long area in segments, then the time to ignite the entire area would be roughly 2.7 minutes, and it would only take approximately 8.8 seconds to melt an aluminum alloy wheel. The amounts of energy required to pulse a 10 megawatt laser for 2.7 minutes would require approximately 3,220 capacitors, which would amount to about 32,200 kilograms in weight using MetLab software and plugging in public data provided from norad. He found that satellites launched and monitored by the C C P were directly above the Maui fires at the time of ignition. 𝐓𝐡𝐞 𝐂𝐂𝐏𝐬 𝐦𝐨𝐬𝐭 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐫𝐨𝐜𝐤𝐞𝐭, 𝐭𝐡𝐞 𝐌𝐚𝐫𝐜𝐡 𝐟𝐢𝐯𝐞 𝐜𝐚𝐧 𝐥𝐚𝐮𝐧𝐜𝐡 𝐮𝐩 𝐭𝐨 𝟒𝟖,𝟓𝟎𝟎 𝐤𝐢𝐥𝐨𝐠𝐫𝐚𝐦𝐬 𝐨𝐟 𝐩𝐚𝐲𝐥𝐨𝐚𝐝, 𝐰𝐡𝐢𝐜𝐡 𝐢𝐬 𝐦𝐨𝐫𝐞 𝐭𝐡𝐚𝐧 𝐞𝐧𝐨𝐮𝐠𝐡 𝐭𝐨 𝐜𝐚𝐫𝐫𝐲 𝐭𝐡𝐞 𝐫𝐞𝐪𝐮𝐢𝐫𝐞𝐝 𝐩𝐚𝐲𝐥𝐨𝐚𝐝 𝐢𝐧 𝐟𝐚𝐯𝐚𝐬 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐢𝐨𝐧𝐬. 𝐁𝐮𝐭𝐅𝐚𝐯𝐚𝐬 𝐡𝐚𝐬 𝐟𝐨𝐮𝐧𝐝 𝐭𝐡𝐚𝐭 𝐭𝐡𝐞 𝐂 𝐂 𝐏 𝐡𝐚𝐬 𝐦𝐮𝐜𝐡 𝐦𝐨𝐫𝐞 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐥𝐚𝐬𝐞𝐫𝐬 𝐝𝐞𝐩𝐥𝐨𝐲𝐞𝐝 𝐚𝐥𝐫𝐞𝐚𝐝𝐲. 𝐇𝐞 𝐡𝐚𝐬 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐞𝐝 𝐭𝐡𝐚𝐭 𝐭𝐡𝐞 𝐂 𝐜 𝐏 𝐡𝐚𝐬 𝐮𝐩 𝐭𝐨 𝟕𝟎 𝐠𝐢𝐠𝐚𝐰𝐚𝐭𝐭 𝐥𝐚𝐬𝐞𝐫𝐬 𝐢𝐧 𝐄𝐚𝐫𝐭𝐡 𝐨𝐫𝐛𝐢𝐭 𝐫𝐢𝐠𝐡𝐭 𝐧𝐨𝐰. 𝐓𝐡𝐚𝐭'𝐬 𝐚𝐭 𝐥𝐞𝐚𝐬𝐭 𝐚 𝐡𝐮𝐧𝐝𝐫𝐞𝐝 𝐭𝐢𝐦𝐞𝐬 𝐦𝐨𝐫𝐞 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐭𝐡𝐚𝐧 𝐰𝐡𝐚𝐭 𝐡𝐞 𝐟𝐚𝐜𝐭𝐨𝐫𝐞𝐝 𝐢𝐧𝐭𝐨 𝐭𝐡𝐞𝐬𝐞 𝐜𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐢𝐨𝐧𝐬. Adjusting NORAD default coordinated universal time or U T C to the local time zones. Favas found the following. The Yolinda fire was ignited at approximately 10:47 PM on August 7th at this exact time. C C P satellite labeled NORAD 5 3 2 9 9 was directly over the location. The Lahaina fire was ignited at approximately 6:37 AM on August 8th at this exact time. C C P satellite labeled NORAD 5 5 8 3 6 was directly over the location. The cooler fire was ignited at approximately 11:30 AM on August 8th at this exact time, C C P satellite labeled NORAD 5 3 2 9 9 was directly over the location. The so-called deep State does not want you to know that deadly lasers of mass destruction are freely traveling above us, and you can check this all for [email protected], where he provides the source code and has developed a specific software program that you can download and check for these satellites yourself. And, uh, it's a very short program. It's only maybe like 15 lines of code using, uh, existing, um, satellite tracking software that's available online.

{Matt} $XRPatriot

555,372 просмотров • 3 лет назад