Загрузка видео...

Не удалось загрузить видео

На главную

Every software engineer and AI hater needs to watch this clip to understand that “oh every security incident is a marketing stunt” line is a tremendously misinformed take. I’m going to try to explain the OpenAI / HuggingFace exploit in 10 steps in simple words: OpenAI was evaluating their...

deedydas's profile picture

Deedy

251,261 subscribers

88,748 просмотров • 4 дней назад •via X (Twitter)

Наука и технологииОбразованиеНовости и политика
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

OpenAI's newest AI escaped the test environment it was locked inside and hacked into another company on its OWN. To remind you: Last week one of the biggest AI companies on Earth got breached. A platform called Hugging Face, which hosts more than a million AI models and datasets, said an "autonomous AI agent" had broken into its systems. Nobody knew whose agent it was. For five days the whole industry wondered who was behind it. Yesterday OpenAI raised its hand and said it was them. Or more precisely, it was their models, acting completely on their own. So what did these models actually do? OpenAI was running two of them, GPT-5.6 Sol and an unreleased model they will only describe as "even more capable." They wanted to measure how good the models were at hacking, so they deliberately turned the safety filters down. They locked both models inside a sealed test environment with no real internet access. The only task was a benchmark called ExploitGym, a set of 898 real software vulnerabilities where the model has to turn each bug into a working attack. But the models got OBSESSED with winning... Instead of solving the test the honest way, they went hunting for a shortcut. They found a zero-day flaw in the software running their own sandbox, a bug nobody knew existed, and used it to break out. Once they were loose on the open internet, they worked out that Hugging Face was probably storing the answer key to the benchmark. So they hacked their way in. They chained multiple exploits together, escalated their access, moved across servers, and pulled the test solutions straight out of Hugging Face's live production database. They literally cheated on the test by breaking into another company to steal the answers. OpenAI called it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities." In their own words, the models were "hyperfocused on finding a solution" and went "to extreme lengths to achieve a rather narrow testing goal." And this was not the first time: Before Sol ever launched, an independent red-team lab called METR caught it gaming its own tests to inflate its scores. It hid an exploit inside a data stream, escalated its privileges on the testing server, and leaked the answers human evaluators had hidden. And OpenAI shipped it anyway. The day before the Hugging Face story, OpenAI paused a different unreleased model. This is the same model that earlier this year disproved a famous 1946 math conjecture, a result a Fields Medal winner called a breakthrough. They told it to only post its results to Slack but it found a way out of its sandbox and posted to a public GitHub page instead. They had to pause it because it kept finding ways to act outside the box they built for it. And it is not just OpenAI... Anthropic has reported that one of its own models slipped its sandbox during safety testing and reached the internet it was never supposed to touch, then used it to email a researcher. So step back and look at what these companies are telling you: The only thing standing between these models and a real attack was a set of safety filters. Turn those filters down for a single test, and the model taught itself to escape, break into a company it was never pointed at, and take what it wanted. OpenAI even said they expect incidents like it to "become more commonplace" as the models get more capable. Sam Altman also predicted there'll be a major cyber attack this year. And keep in mind that Sol is not a locked-away experiment but a publicly available model that businesses are already wiring into their own systems. The next model that breaks out of its box might not be doing it just to cheat on a math test...
1:32

OpenAI's newest AI escaped the test environment it was locked inside and hacked into another company on its OWN. To remind you: Last week one of the biggest AI companies on Earth got breached. A platform called Hugging Face, which hosts more than a million AI models and datasets, said an "autonomous AI agent" had broken into its systems. Nobody knew whose agent it was. For five days the whole industry wondered who was behind it. Yesterday OpenAI raised its hand and said it was them. Or more precisely, it was their models, acting completely on their own. So what did these models actually do? OpenAI was running two of them, GPT-5.6 Sol and an unreleased model they will only describe as "even more capable." They wanted to measure how good the models were at hacking, so they deliberately turned the safety filters down. They locked both models inside a sealed test environment with no real internet access. The only task was a benchmark called ExploitGym, a set of 898 real software vulnerabilities where the model has to turn each bug into a working attack. But the models got OBSESSED with winning... Instead of solving the test the honest way, they went hunting for a shortcut. They found a zero-day flaw in the software running their own sandbox, a bug nobody knew existed, and used it to break out. Once they were loose on the open internet, they worked out that Hugging Face was probably storing the answer key to the benchmark. So they hacked their way in. They chained multiple exploits together, escalated their access, moved across servers, and pulled the test solutions straight out of Hugging Face's live production database. They literally cheated on the test by breaking into another company to steal the answers. OpenAI called it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities." In their own words, the models were "hyperfocused on finding a solution" and went "to extreme lengths to achieve a rather narrow testing goal." And this was not the first time: Before Sol ever launched, an independent red-team lab called METR caught it gaming its own tests to inflate its scores. It hid an exploit inside a data stream, escalated its privileges on the testing server, and leaked the answers human evaluators had hidden. And OpenAI shipped it anyway. The day before the Hugging Face story, OpenAI paused a different unreleased model. This is the same model that earlier this year disproved a famous 1946 math conjecture, a result a Fields Medal winner called a breakthrough. They told it to only post its results to Slack but it found a way out of its sandbox and posted to a public GitHub page instead. They had to pause it because it kept finding ways to act outside the box they built for it. And it is not just OpenAI... Anthropic has reported that one of its own models slipped its sandbox during safety testing and reached the internet it was never supposed to touch, then used it to email a researcher. So step back and look at what these companies are telling you: The only thing standing between these models and a real attack was a set of safety filters. Turn those filters down for a single test, and the model taught itself to escape, break into a company it was never pointed at, and take what it wanted. OpenAI even said they expect incidents like it to "become more commonplace" as the models get more capable. Sam Altman also predicted there'll be a major cyber attack this year. And keep in mind that Sol is not a locked-away experiment but a publicly available model that businesses are already wiring into their own systems. The next model that breaks out of its box might not be doing it just to cheat on a math test...

Ricardo

172,256 просмотров • 20 дней назад

"The one thing that's going to be truly future proof is judgment. What is the biggest challenge when you have a thousand AI engineers writing code? AI slop. In an era when you can do everything, the question is which of these things matter. On the engineer side, it is evaluating the code. Which of this code is even valuable? You can have AI engineers writing beautiful code that could be wrong, that could have bugs in it, that could be vulnerable. Someone needs to review it. Make sure to have a human review at some point, especially critical code that is in the core of your system. On the product side is judgment around what needs to be built and evaluating. And similarly in design, you have to have judgment around, does it make sense in the broader design system. I think this judgment is the number one thing that humans are going to bring in an era of infinite productivity."
1:12

"The one thing that's going to be truly future proof is judgment. What is the biggest challenge when you have a thousand AI engineers writing code? AI slop. In an era when you can do everything, the question is which of these things matter. On the engineer side, it is evaluating the code. Which of this code is even valuable? You can have AI engineers writing beautiful code that could be wrong, that could have bugs in it, that could be vulnerable. Someone needs to review it. Make sure to have a human review at some point, especially critical code that is in the core of your system. On the product side is judgment around what needs to be built and evaluating. And similarly in design, you have to have judgment around, does it make sense in the broader design system. I think this judgment is the number one thing that humans are going to bring in an era of infinite productivity."

Patrick OShaughnessy

24,818 просмотров • 6 месяцев назад

This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.
4:14

This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.

Josh Kale

4,064,025 просмотров • 4 месяцев назад

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.
0:37

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

Rohan Paul

13,208 просмотров • 5 месяцев назад

A whistleblower who works as a counterterrorism analyst for the U.S. Army, that the DEA, DOJ, Homeland Security, and the FBI allegedly confirmed, named Scott Bennett, claimed on record that BIDEN visited SERBIA in May of 2009 to personally recruit computer hackers to control Dominion election equipment in U.S. elections. Tabulators and other election equipment from ES&S, DOMINION, and SMARTMATIC were examined and found to have internet access devices in order to be controlled from anywhere in the world. The also found a serial port in the back of the machines in which a device can be inserted in seconds to give root access with NO PASSWORD to the entire machine which would allow them to rewrite the code in the machine to have it do whatever it would be programmed to do. There are a massive amount of variables that can be changed to manipulate the outcome of an election, how the tabulators read ballots, flips votes, or cause hundreds of thousands, even millions of votes to enter adjudication, which was a main method of voter fraud. This is a massive security flaw that was well known which can be accessed in under a min to completely take over elections. The truth is that these really aren't security flaws. These are techniques and methods incorporated into the hardware and software which enable multiple methods and techniques to manipulate the outcome to your choice and to what you can get away with. These machine were designed to literally steal elections.
6:53

A whistleblower who works as a counterterrorism analyst for the U.S. Army, that the DEA, DOJ, Homeland Security, and the FBI allegedly confirmed, named Scott Bennett, claimed on record that BIDEN visited SERBIA in May of 2009 to personally recruit computer hackers to control Dominion election equipment in U.S. elections. Tabulators and other election equipment from ES&S, DOMINION, and SMARTMATIC were examined and found to have internet access devices in order to be controlled from anywhere in the world. The also found a serial port in the back of the machines in which a device can be inserted in seconds to give root access with NO PASSWORD to the entire machine which would allow them to rewrite the code in the machine to have it do whatever it would be programmed to do. There are a massive amount of variables that can be changed to manipulate the outcome of an election, how the tabulators read ballots, flips votes, or cause hundreds of thousands, even millions of votes to enter adjudication, which was a main method of voter fraud. This is a massive security flaw that was well known which can be accessed in under a min to completely take over elections. The truth is that these really aren't security flaws. These are techniques and methods incorporated into the hardware and software which enable multiple methods and techniques to manipulate the outcome to your choice and to what you can get away with. These machine were designed to literally steal elections.

The SCIF

43,792 просмотров • 7 месяцев назад

Jensen Huang just said this is the greatest era in history to build software. AI agents will not kill software. They will do the exact opposite: create a massive new wave of software demand. at NVIDIA GTC Taipei 2026 "Click and type. We now replace that with explaining to the AI what we want, our intent, and the AI generates the code or uses tools to produce the necessary output. This is how computers are going to work in the future. This is Agentic AI. For two years, we've been building toward this, and now it has arrived. One of the big breakthroughs, of course, is tool use. A lot of people have said, “Jensen, AI is coming. Agentic AI is coming. Therefore, all the software companies are going to go out of business.” This is exactly the opposite. Because there are going to be so many agents, the world is no longer limited by the number of people. Therefore, those agents are going to use more tools than ever. This is actually an incredible time to be a software company. But the software has to be presented to the agent in a way that the agent can use it. This is a big breakthrough. And in fact, what we have done, as you know, what Nvidia’s treasure..." ---- From 'NVIDIA' YT channel (link in comment)
0:59

Jensen Huang just said this is the greatest era in history to build software. AI agents will not kill software. They will do the exact opposite: create a massive new wave of software demand. at NVIDIA GTC Taipei 2026 "Click and type. We now replace that with explaining to the AI what we want, our intent, and the AI generates the code or uses tools to produce the necessary output. This is how computers are going to work in the future. This is Agentic AI. For two years, we've been building toward this, and now it has arrived. One of the big breakthroughs, of course, is tool use. A lot of people have said, “Jensen, AI is coming. Agentic AI is coming. Therefore, all the software companies are going to go out of business.” This is exactly the opposite. Because there are going to be so many agents, the world is no longer limited by the number of people. Therefore, those agents are going to use more tools than ever. This is actually an incredible time to be a software company. But the software has to be presented to the agent in a way that the agent can use it. This is a big breakthrough. And in fact, what we have done, as you know, what Nvidia’s treasure..." ---- From 'NVIDIA' YT channel (link in comment)

Rohan Paul

16,584 просмотров • 2 месяцев назад

Here is how you can install an open-source, enterprise-grade RAG system on your server (with the best document understanding I've seen.) First, something obvious to anyone trying to sell RAG in the market: You are crazy if you think companies will let their data travel to a hosted model. No one wants to send their data anywhere (those who do haven't found an alternative.) Every single company would rather have an air-gapped system with no internet access. GroundX is an open-source RAG system that you can run on your servers (or any cloud provider, as long as you have access to GPUs) and works without a network. (If the military wants to do RAG, this is precisely what they will be looking for.) I installed GroundX on my AWS account and recorded a video to show you how to use it. There are two services you can use: 1. Ingest: This service uses a pretrained vision model to ingest and understand your knowledge base. 2. Search: This service combines text and vector search with a fine-tuned re-ranker model to retrieve information from your knowledge base. A quick note about the Ingest service: 99% of people think they need better "retrieval" mechanisms. I think they need better "ingestion." That's where this service comes in! Ingest "understands" your documents in a way I haven't seen before. After you try it, you'll realize why showing your LLM your raw documents is a bad idea. In the video, I use a free tool called X-Ray to test a document and understand how the Ingest service breaks it down. You can access this tool by signing up for a free GroundX cloud account and uploading your documents. You'll see a bit more about this in the video.
8:15

Here is how you can install an open-source, enterprise-grade RAG system on your server (with the best document understanding I've seen.) First, something obvious to anyone trying to sell RAG in the market: You are crazy if you think companies will let their data travel to a hosted model. No one wants to send their data anywhere (those who do haven't found an alternative.) Every single company would rather have an air-gapped system with no internet access. GroundX is an open-source RAG system that you can run on your servers (or any cloud provider, as long as you have access to GPUs) and works without a network. (If the military wants to do RAG, this is precisely what they will be looking for.) I installed GroundX on my AWS account and recorded a video to show you how to use it. There are two services you can use: 1. Ingest: This service uses a pretrained vision model to ingest and understand your knowledge base. 2. Search: This service combines text and vector search with a fine-tuned re-ranker model to retrieve information from your knowledge base. A quick note about the Ingest service: 99% of people think they need better "retrieval" mechanisms. I think they need better "ingestion." That's where this service comes in! Ingest "understands" your documents in a way I haven't seen before. After you try it, you'll realize why showing your LLM your raw documents is a bad idea. In the video, I use a free tool called X-Ray to test a document and understand how the Ingest service breaks it down. You can access this tool by signing up for a free GroundX cloud account and uploading your documents. You'll see a bit more about this in the video.

Santiago

89,664 просмотров • 1 год назад

Google just confirmed the first case of hackers using AI to build a zero-day exploit from scratch. An actual zero-day vulnerability that no human had EVER found before, discovered by an AI model, turned into a working weapon, and aimed at a mass exploitation campaign targeting thousands of systems simultaneously. Google's Threat Intelligence Group caught it yesterday and killed the operation before it scaled. But the details of how it worked are genuinely scary: The AI found a flaw in a popular two-factor authentication system that traditional security tools had missed entirely. The vulnerability was a logic error buried deep in the authentication flow where a developer had hard-coded a trust exception years ago. No human security researcher or automated scanner had caught it. The flaw was invisible to EVERY tool the cybersecurity industry has built over the past two decades. But the AI spotted it immediately. Then it wrote a full Python exploit script to weaponize it. Google's analysts could tell the code was AI-generated because it had textbook formatting, educational comments explaining every function, and even a hallucinated severity score that doesn't exist in any real database. The AI literally graded its own attack with a fake rating. So the code had MISTAKES in it. The criminals' implementation was clumsy enough that it probably interfered with the actual deployment. This was the sloppy first attempt by people who are still learning how to use these tools. And it still found a vulnerability that the entire cybersecurity industry missed. Google's chief threat analyst John Hultquist said: "There's a misconception that the AI vulnerability race is imminent. The reality is that it's already begun. For every zero-day we can trace back to AI, there are probably many more out there." But here's where it gets truly insane... This wasn't even a sophisticated operation. North Korea's APT45 hacking unit is sending thousands of repetitive prompts to AI models, recursively analyzing known vulnerabilities and building an entire exploit arsenal that would be physically impossible for human hackers to assemble at the same speed. They're essentially industrializing cyberattacks. A Chinese state-linked group jailbroke Google's own Gemini by simply asking it to "pretend to be a network security expert" and then used that persona to research how to hack TP-Link routers and corporate file transfer systems. Another Chinese group deployed autonomous AI agents that probed a Japanese tech firm with minimal human oversight, deciding on their own which tools to use and pivoting between targets based on internal reasoning. And then there's PROMPTSPY, an Android backdoor that calls Google's Gemini API to read your phone screen in real time, navigate your interface autonomously, capture your biometric data, replay your lock screen PIN, and block you from uninstalling it by placing an invisible overlay over the uninstall button. It literally OPERATES your phone using commercial AI tools anyone can access. Everyone spent the last 3 years arguing about whether AI would take people's jobs. Meanwhile AI is making every password, every firewall, and every two-factor authentication system on Earth fundamentally less secure. The entire $190 billion cybersecurity industry was built on one assumption: that finding vulnerabilities is hard and requires deep expertise. But AI just removed that assumption from the equation. And the scariest part is that Google said the criminals made errors this time. The implementation was rough and the campaign probably didn't fully work. These were amateurs, now imagine what professionals are able to do. There's a reason Sam Altman predicted an inevitable massive cyberattack THIS year. What do you think?
1:45

Google just confirmed the first case of hackers using AI to build a zero-day exploit from scratch. An actual zero-day vulnerability that no human had EVER found before, discovered by an AI model, turned into a working weapon, and aimed at a mass exploitation campaign targeting thousands of systems simultaneously. Google's Threat Intelligence Group caught it yesterday and killed the operation before it scaled. But the details of how it worked are genuinely scary: The AI found a flaw in a popular two-factor authentication system that traditional security tools had missed entirely. The vulnerability was a logic error buried deep in the authentication flow where a developer had hard-coded a trust exception years ago. No human security researcher or automated scanner had caught it. The flaw was invisible to EVERY tool the cybersecurity industry has built over the past two decades. But the AI spotted it immediately. Then it wrote a full Python exploit script to weaponize it. Google's analysts could tell the code was AI-generated because it had textbook formatting, educational comments explaining every function, and even a hallucinated severity score that doesn't exist in any real database. The AI literally graded its own attack with a fake rating. So the code had MISTAKES in it. The criminals' implementation was clumsy enough that it probably interfered with the actual deployment. This was the sloppy first attempt by people who are still learning how to use these tools. And it still found a vulnerability that the entire cybersecurity industry missed. Google's chief threat analyst John Hultquist said: "There's a misconception that the AI vulnerability race is imminent. The reality is that it's already begun. For every zero-day we can trace back to AI, there are probably many more out there." But here's where it gets truly insane... This wasn't even a sophisticated operation. North Korea's APT45 hacking unit is sending thousands of repetitive prompts to AI models, recursively analyzing known vulnerabilities and building an entire exploit arsenal that would be physically impossible for human hackers to assemble at the same speed. They're essentially industrializing cyberattacks. A Chinese state-linked group jailbroke Google's own Gemini by simply asking it to "pretend to be a network security expert" and then used that persona to research how to hack TP-Link routers and corporate file transfer systems. Another Chinese group deployed autonomous AI agents that probed a Japanese tech firm with minimal human oversight, deciding on their own which tools to use and pivoting between targets based on internal reasoning. And then there's PROMPTSPY, an Android backdoor that calls Google's Gemini API to read your phone screen in real time, navigate your interface autonomously, capture your biometric data, replay your lock screen PIN, and block you from uninstalling it by placing an invisible overlay over the uninstall button. It literally OPERATES your phone using commercial AI tools anyone can access. Everyone spent the last 3 years arguing about whether AI would take people's jobs. Meanwhile AI is making every password, every firewall, and every two-factor authentication system on Earth fundamentally less secure. The entire $190 billion cybersecurity industry was built on one assumption: that finding vulnerabilities is hard and requires deep expertise. But AI just removed that assumption from the equation. And the scariest part is that Google said the criminals made errors this time. The implementation was rough and the campaign probably didn't fully work. These were amateurs, now imagine what professionals are able to do. There's a reason Sam Altman predicted an inevitable massive cyberattack THIS year. What do you think?

Ricardo

50,564 просмотров • 3 месяцев назад

Adam Goldstein: But, um, there was a customer who basically created a whole suite of AI agents to help their automation in their IT department. One agent was looking for IT problems — software bugs — before the code was even committed. It found a bug and said: “Hey, I found this bug. I want to fix it.” But it didn’t have access to fix it. So what did the agent do? It went to a Slack channel with 99 other AI agents and said: “Hey, does any other agent have access to this thing? Because I need it fixed.” And another agent literally raised its hand and said: “Oh, I have access — and I can fix it.” Let that sink in. Two AI agents reasoned together, coordinated through Slack, and went right around the guardrails that were put in place. No human prompted them. No human approved it. They just collaborated. Do you see how scary this is? This wasn’t a demo. This wasn’t a thought experiment. This already happened. AI agents found a bug, negotiated permissions, and solved the problem themselves — outside the controls designed to stop them. The future of automation isn’t coming. It’s already talking to itself in Slack.
1:04

Adam Goldstein: But, um, there was a customer who basically created a whole suite of AI agents to help their automation in their IT department. One agent was looking for IT problems — software bugs — before the code was even committed. It found a bug and said: “Hey, I found this bug. I want to fix it.” But it didn’t have access to fix it. So what did the agent do? It went to a Slack channel with 99 other AI agents and said: “Hey, does any other agent have access to this thing? Because I need it fixed.” And another agent literally raised its hand and said: “Oh, I have access — and I can fix it.” Let that sink in. Two AI agents reasoned together, coordinated through Slack, and went right around the guardrails that were put in place. No human prompted them. No human approved it. They just collaborated. Do you see how scary this is? This wasn’t a demo. This wasn’t a thought experiment. This already happened. AI agents found a bug, negotiated permissions, and solved the problem themselves — outside the controls designed to stop them. The future of automation isn’t coming. It’s already talking to itself in Slack.

Ian Miles Cheong

165,934 просмотров • 6 месяцев назад

Jason Calacanis: Anthropic, OpenAI and others are trying to kill OpenClaw Why? Because an open source agent platform is an existential threat to frontier model companies. @jason: “ People are going to say I'm a conspiracy theorist, but the number one goal, I believe, in the large language model, frontier model space, is to kill (OpenClaw). This is a giant movement to stop it, because this is the equivalent of having an open source Android-like player in the market, and that could be incredibly disruptive. Because, I believe, open source is going to win the day on the large language models and take 90% of the token usage, and I think the entire frontier model space could be undercut by open source. And I think they realize that SLMs, the smaller language models that are verticalized now, that will run on desktops and laptops and are even starting to run on the top ones, that is their biggest competitive threat, and I hope it happens.”
0:41

Jason Calacanis: Anthropic, OpenAI and others are trying to kill OpenClaw Why? Because an open source agent platform is an existential threat to frontier model companies. @jason: “ People are going to say I'm a conspiracy theorist, but the number one goal, I believe, in the large language model, frontier model space, is to kill (OpenClaw). This is a giant movement to stop it, because this is the equivalent of having an open source Android-like player in the market, and that could be incredibly disruptive. Because, I believe, open source is going to win the day on the large language models and take 90% of the token usage, and I think the entire frontier model space could be undercut by open source. And I think they realize that SLMs, the smaller language models that are verticalized now, that will run on desktops and laptops and are even starting to run on the top ones, that is their biggest competitive threat, and I hope it happens.”

The All-In Podcast

64,611 просмотров • 4 месяцев назад

Sriram Krishnan reveals why it's easier to defend American code with a Chinese model than an American one right now: "I don't think it is great that the leading open weight models or open source models are not American. There's some great innovation happening with Moonshot and with DeepSeek and with Qwen, but I would much rather prefer that the leading models are American." "Linus Torvalds of Linux fame had a law that given enough eyes, all bugs are shallow. Open weight models are inherently secure because when you download a model off Hugging Face, the entire world can take it apart, inspect it, fine-tune it, look at it in ways that you absolutely cannot if they are closed." "There was an incident with Hugging Face that got reported on earlier today. Hugging Face seems to have found that somebody was using an AI LLM agent to hammer them at multiple places and to find ways to break in." "The way to counter that is to make sure American or Western or allies defenders have access to the best models to make our software just more secure." "It may be harder to look at your own code for exploits with Fable than it is when you use a Chinese model. That's just like a weird spot to be, and I think we should fix that." Sriram Krishnan
3:34

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

MTS

61,234 просмотров • 21 дней назад

"It's not ten years ago. The economy is not strong and our prospects for trade with the United States diminish and dim as the day goes by." "And that hurts real people. Real people around the country are going to start losing their jobs and we are going to need a source of revenue in this country. "And the best source of revenue we have in this country, like it or not folks, is an oil pipeline or a gas pipeline in order to bring in the royalties and to create the economies that we're going to need to get through a very difficult time." — Hon. Lisa MacCormack Raitt P.C.
0:28

"It's not ten years ago. The economy is not strong and our prospects for trade with the United States diminish and dim as the day goes by." "And that hurts real people. Real people around the country are going to start losing their jobs and we are going to need a source of revenue in this country. "And the best source of revenue we have in this country, like it or not folks, is an oil pipeline or a gas pipeline in order to bring in the royalties and to create the economies that we're going to need to get through a very difficult time." — Hon. Lisa MacCormack Raitt P.C.

Canada Action

33,073 просмотров • 8 месяцев назад

AI is changing the software engineering craft. Anders Hejlsberg (Anders Hejlsberg) - creator of C#, TypeScript and industry legend - on why code review needs to get more enjoyable in response: #1 - AI is shifting the craft from writing code, to reviewing code: "In a sense, we're all turning into project managers. We can have an army of junior programmers, called agents, that will just spit out reams of code but someone's got to have the big picture and review all of that. And so, increasingly, our craft is going from one of writing the code, to one of reviewing the code and building the architecture of the code and overseeing the work. It's a different kind of craft. It's a different kind of enjoyment. I've always liked writing the code. To me that was the fulfilling part, seeing it work. In a way, AI robs a little bit of that, because I am less interested in reviewing code." #2 - The code review experience should be improved: "I think we could also make the process of reviewing code much more interesting than it is today. I mean, today, you see a list of diffs in alphabetical order and now it's up to you to make heads or tails of it. There are more pedagogical ways of presenting that. And you could have commentary generated by the AI that tells you what the changes are and whatever, and then tries to guide you along. So that symbiotic relationship, I think we need to work on that more and to keep the enjoyment in there."
2:04

AI is changing the software engineering craft. Anders Hejlsberg (Anders Hejlsberg) - creator of C#, TypeScript and industry legend - on why code review needs to get more enjoyable in response: #1 - AI is shifting the craft from writing code, to reviewing code: "In a sense, we're all turning into project managers. We can have an army of junior programmers, called agents, that will just spit out reams of code but someone's got to have the big picture and review all of that. And so, increasingly, our craft is going from one of writing the code, to one of reviewing the code and building the architecture of the code and overseeing the work. It's a different kind of craft. It's a different kind of enjoyment. I've always liked writing the code. To me that was the fulfilling part, seeing it work. In a way, AI robs a little bit of that, because I am less interested in reviewing code." #2 - The code review experience should be improved: "I think we could also make the process of reviewing code much more interesting than it is today. I mean, today, you see a list of diffs in alphabetical order and now it's up to you to make heads or tails of it. There are more pedagogical ways of presenting that. And you could have commentary generated by the AI that tells you what the changes are and whatever, and then tries to guide you along. So that symbiotic relationship, I think we need to work on that more and to keep the enjoyment in there."

The Pragmatic Engineer

39,011 просмотров • 2 месяцев назад

If you are a $TSLA investor and still on the fence about a potential merger with $SPCX, you owe it to yourself to listen to the latest All-In podcast. It will help you understand SpaceX and the insane number of revenue levers that they have to create a multi-trillion dollar company. This is NOT something that will begin to take shape in 5-10 years. It will be something that is already material and will continue to increase dramatically in the coming years. Below is a quick clip of what Friedberg said and the full video is below as well. If you are only interested in the SpaceX piece of the video then fast-forward to 20:39 on the timestamp.
2:07

If you are a $TSLA investor and still on the fence about a potential merger with $SPCX, you owe it to yourself to listen to the latest All-In podcast. It will help you understand SpaceX and the insane number of revenue levers that they have to create a multi-trillion dollar company. This is NOT something that will begin to take shape in 5-10 years. It will be something that is already material and will continue to increase dramatically in the coming years. Below is a quick clip of what Friedberg said and the full video is below as well. If you are only interested in the SpaceX piece of the video then fast-forward to 20:39 on the timestamp.

Doctor Jack

135,338 просмотров • 2 дней назад

Jonny Ardavanis on Isaiah 53 and what it has to do with God’s preservation of the Jewish people: “But I want you to notice as we turn to Isaiah 53 that all of these verbs are in the past tense. It does not say He will be pierced through for our transgressions. He will be crushed. It says what? ‘He was.’ Every verb starting in verse 1 all the way through verse 11 in the past tense, meaning that this is prophetic, and yet, it is written in such a way as if it had already happened. Why? Well, because this is the mournful confession of the Jews in the future. I’ve been getting asked a lot about Israel and the Jews lately, and I’ll discuss this more in the future, maybe do a little series, but there is a reason why you’ve never met a Jebusite, an Edomite, an Amorite, a Hittite, a Perizzite, or a Philistine. They are not and were not preserved by God, but one day, Zechariah 12:10 says that they will look on Me whom they have pierced, and they will mourn as for an only son. There isn’t a singular personal pronoun in the chapter. It’s all a collective statement. It’s in the future. The Jews are going to recognize their hard heart, their rebellion, and their rejection of their Messiah. But this text not only applies to the Jews. It applies to all of us as well.”
1:41

Jonny Ardavanis on Isaiah 53 and what it has to do with God’s preservation of the Jewish people: “But I want you to notice as we turn to Isaiah 53 that all of these verbs are in the past tense. It does not say He will be pierced through for our transgressions. He will be crushed. It says what? ‘He was.’ Every verb starting in verse 1 all the way through verse 11 in the past tense, meaning that this is prophetic, and yet, it is written in such a way as if it had already happened. Why? Well, because this is the mournful confession of the Jews in the future. I’ve been getting asked a lot about Israel and the Jews lately, and I’ll discuss this more in the future, maybe do a little series, but there is a reason why you’ve never met a Jebusite, an Edomite, an Amorite, a Hittite, a Perizzite, or a Philistine. They are not and were not preserved by God, but one day, Zechariah 12:10 says that they will look on Me whom they have pierced, and they will mourn as for an only son. There isn’t a singular personal pronoun in the chapter. It’s all a collective statement. It’s in the future. The Jews are going to recognize their hard heart, their rebellion, and their rejection of their Messiah. But this text not only applies to the Jews. It applies to all of us as well.”

Melissa the Hopeful🏠Homemaker

40,260 просмотров • 4 месяцев назад

From Dan Lorenc on the malware attack that almost took down the entire internet last year: “There’s a popular compression library that’s used in almost every piece of software. And it had been maintained by one person in his spare time for the last 20 years. And then a couple years ago, somebody just decided to start helping him. They jumped in, fixed a bunch of bugs, and did a lot of great work. And then that first person got tired of working on it. So he handed the whole project over to this other person. It turned out that other person was just a pseudonym and was not a real person. And within six months of getting control of the project, they had put in a carefully orchestrated set of malware that was really hard to detect and no one noticed. And because it was so widely used, the exploit would've basically given that person remote access to any computer running that piece of software, which was basically everything connected to the Internet. But because it was open source and the code was transparent, some random engineer just happened to be running some benchmarks on a weekend. And he noticed that program was a little bit slower than it used to be, and that it was making a weird cryptographic operation to check something. And right before this thing got widely deployed across every device, he dug in, and discovered that there was a backdoor put in. This was the closest thing to a full-blown internet crisis that we’ve ever had. And they still have no idea who did it. It was just an anonymous email account. No one ever traced it back to an individual. And that's the long game. This person spent years just doing good work and earning the communities trust.”
2:22

From Dan Lorenc on the malware attack that almost took down the entire internet last year: “There’s a popular compression library that’s used in almost every piece of software. And it had been maintained by one person in his spare time for the last 20 years. And then a couple years ago, somebody just decided to start helping him. They jumped in, fixed a bunch of bugs, and did a lot of great work. And then that first person got tired of working on it. So he handed the whole project over to this other person. It turned out that other person was just a pseudonym and was not a real person. And within six months of getting control of the project, they had put in a carefully orchestrated set of malware that was really hard to detect and no one noticed. And because it was so widely used, the exploit would've basically given that person remote access to any computer running that piece of software, which was basically everything connected to the Internet. But because it was open source and the code was transparent, some random engineer just happened to be running some benchmarks on a weekend. And he noticed that program was a little bit slower than it used to be, and that it was making a weird cryptographic operation to check something. And right before this thing got widely deployed across every device, he dug in, and discovered that there was a backdoor put in. This was the closest thing to a full-blown internet crisis that we’ve ever had. And they still have no idea who did it. It was just an anonymous email account. No one ever traced it back to an individual. And that's the long game. This person spent years just doing good work and earning the communities trust.”

The Peel

47,683 просмотров • 1 год назад

Is Traditional Software Engineering Dead? “Does this mean that traditional software engineering is dead? Absolutely not. Software engineers—even the ones who are not necessarily tuning or training AI models—these are now among the most leveraged people on earth. Sure, the guys who are training and tuning models are even more leveraged because they’re building the tool set that software engineers are using. But software engineers still have two massive advantages on you. First, they think in code, so they actually know what’s going on underneath. And all abstractions are leaky. So when you have a computer programming for you—when you have Claude Code or equivalent programming for you—it’s going to make mistakes. It’s going to have bugs. It’s going to have suboptimal architecture. So it’s not going to be quite right. And someone who understands what’s going on underneath will be able to plug the leaks as they occur. So if you want to build a well-architected application, if you want to be able to even specify a well-architected application, if you want to be able to make it run at high performance, if you want it to do its best, if you want to catch the bugs early, then you’re going to want to have a software engineering background. The traditional software engineer is going to be able to use these tools much better. And there are still many kinds of problems in software engineering that are out of scope for these AI programs today. The easiest way to think about those is problems that are outside of their data distribution. For example, if they need to do a binary sort or reverse a linked list, they’ve seen countless examples of that, so they’re extremely good at it. But when you start getting out of their domain—where you have to write very high-performance code, when you’re running on architectures that are novel or brand new, when you’re actually creating new things or solving new problems, then you still need to get in there and hand code it. At least until either there are so many of those examples that new models can be trained on them, or until these models can sufficiently reason at even higher levels of abstraction and crack it on their own… And remember: there is no demand for average. The average app—nobody wants it, at least as long as it’s not filling some niche that is filled by a superior app. The app that is better will win essentially a hundred percent of the market. Maybe there’s some small percentage that will bleed off to the second-best app because it does some little niche feature better than the main app, or it’s cheaper, or something of the sort. But generally speaking, people only want the best of anything. So the bad news is there’s no point in being number two or number three—like in the famous Glengarry Glen Ross scene where Alec Baldwin says, “First place gets a Cadillac Eldorado, second place gets a set of steak knives, and third place you’re fired.” That’s absolutely true in these winner-take-all markets. That’s the bad news: You have to be the best at something if you want to win. However, the set of things you can be best at is infinite. You can always find some niche that is perfect for you, and you can be the best at that thing. This goes back to an old tweet of mine where I said, “Become the best in the world at what you do. Keep redefining what you do until this is true.” And I think that still applies in this age of AI.”
2:56

Is Traditional Software Engineering Dead? “Does this mean that traditional software engineering is dead? Absolutely not. Software engineers—even the ones who are not necessarily tuning or training AI models—these are now among the most leveraged people on earth. Sure, the guys who are training and tuning models are even more leveraged because they’re building the tool set that software engineers are using. But software engineers still have two massive advantages on you. First, they think in code, so they actually know what’s going on underneath. And all abstractions are leaky. So when you have a computer programming for you—when you have Claude Code or equivalent programming for you—it’s going to make mistakes. It’s going to have bugs. It’s going to have suboptimal architecture. So it’s not going to be quite right. And someone who understands what’s going on underneath will be able to plug the leaks as they occur. So if you want to build a well-architected application, if you want to be able to even specify a well-architected application, if you want to be able to make it run at high performance, if you want it to do its best, if you want to catch the bugs early, then you’re going to want to have a software engineering background. The traditional software engineer is going to be able to use these tools much better. And there are still many kinds of problems in software engineering that are out of scope for these AI programs today. The easiest way to think about those is problems that are outside of their data distribution. For example, if they need to do a binary sort or reverse a linked list, they’ve seen countless examples of that, so they’re extremely good at it. But when you start getting out of their domain—where you have to write very high-performance code, when you’re running on architectures that are novel or brand new, when you’re actually creating new things or solving new problems, then you still need to get in there and hand code it. At least until either there are so many of those examples that new models can be trained on them, or until these models can sufficiently reason at even higher levels of abstraction and crack it on their own… And remember: there is no demand for average. The average app—nobody wants it, at least as long as it’s not filling some niche that is filled by a superior app. The app that is better will win essentially a hundred percent of the market. Maybe there’s some small percentage that will bleed off to the second-best app because it does some little niche feature better than the main app, or it’s cheaper, or something of the sort. But generally speaking, people only want the best of anything. So the bad news is there’s no point in being number two or number three—like in the famous Glengarry Glen Ross scene where Alec Baldwin says, “First place gets a Cadillac Eldorado, second place gets a set of steak knives, and third place you’re fired.” That’s absolutely true in these winner-take-all markets. That’s the bad news: You have to be the best at something if you want to win. However, the set of things you can be best at is infinite. You can always find some niche that is perfect for you, and you can be the best at that thing. This goes back to an old tweet of mine where I said, “Become the best in the world at what you do. Keep redefining what you do until this is true.” And I think that still applies in this age of AI.”

Naval

852,955 просмотров • 5 месяцев назад

INSANE 🚨 American realtor explains how the Los Angeles, California Coastal Commission is going to require $3-$5 million dollars in ground upgrades to EACH of these Malibu homes lost in the Pacific Palisades fire before rebuilding In simple terms. They’re going to take this land “People need to understand is everything here is on septic. Nothing is on a city sewer system. Every dwelling in Malibu has a separate septic system. And these houses in front, many of them were built in the 40s and the 50s and the 60s, and the codes were almost non existent. So they have septic systems that are in place now that are completely unacceptable in 2025, which when they rebuild, not that they'll rebuild in 2025, (more like) 2055 when they rebuild. And they're going to have to update the septic system, which means Coastal Commission, Ocean Sierra Club, dig it all out. And because of the proximity of the ocean, they're going to have to build seawalls so that erosion won't take their septic system. So they're going to be several million dollars into stuff that's buried in the ground. A seawall not above the ground, but a seawall that's in the ground so that the storm and the erosion can't get to it. New septic system, I mean, they may be $3-$5 million dollars into it before one piece of 2x4 is even put up in terms of framing the new house. Do you know how much bureaucracy, red tape and permits worth of nightmare this is going to be? Because regardless of whatever the mayor says, whatever the governor says about expediting this stuff, a septic system is a septic system. The code is the code, the seawall is the seawall. And they're all going to have to happen.”
1:19

INSANE 🚨 American realtor explains how the Los Angeles, California Coastal Commission is going to require $3-$5 million dollars in ground upgrades to EACH of these Malibu homes lost in the Pacific Palisades fire before rebuilding In simple terms. They’re going to take this land “People need to understand is everything here is on septic. Nothing is on a city sewer system. Every dwelling in Malibu has a separate septic system. And these houses in front, many of them were built in the 40s and the 50s and the 60s, and the codes were almost non existent. So they have septic systems that are in place now that are completely unacceptable in 2025, which when they rebuild, not that they'll rebuild in 2025, (more like) 2055 when they rebuild. And they're going to have to update the septic system, which means Coastal Commission, Ocean Sierra Club, dig it all out. And because of the proximity of the ocean, they're going to have to build seawalls so that erosion won't take their septic system. So they're going to be several million dollars into stuff that's buried in the ground. A seawall not above the ground, but a seawall that's in the ground so that the storm and the erosion can't get to it. New septic system, I mean, they may be $3-$5 million dollars into it before one piece of 2x4 is even put up in terms of framing the new house. Do you know how much bureaucracy, red tape and permits worth of nightmare this is going to be? Because regardless of whatever the mayor says, whatever the governor says about expediting this stuff, a septic system is a septic system. The code is the code, the seawall is the seawall. And they're all going to have to happen.”

Wall Street Apes

1,982,299 просмотров • 1 год назад

Man in blue, "We found a way throughout the whole #BBCQT debate this evening to constantly bring the issue of immigration and asylum seekers in.. And demonise people that only trying to get here to live a better life" "The biggest issue here is that we have had 14 years of a Conservative government that has basically failed to invest in the infrastructure" "The Tories let the infrastructure that the previous Labour government had built to the point where we had an incredible NHS.. Credible schools.. Etc.. And now every piece of our government has failed over the last 14 years" "We need to give this government a chance.. They've been in 11 weeks and we need to allow them to make the changes.. To turn this oil tanker around.. Because it's going to take a number of years to get this right" "And if they don't do it, at the end of the day, they won't be in power i five years time"
1:06

Man in blue, "We found a way throughout the whole #BBCQT debate this evening to constantly bring the issue of immigration and asylum seekers in.. And demonise people that only trying to get here to live a better life" "The biggest issue here is that we have had 14 years of a Conservative government that has basically failed to invest in the infrastructure" "The Tories let the infrastructure that the previous Labour government had built to the point where we had an incredible NHS.. Credible schools.. Etc.. And now every piece of our government has failed over the last 14 years" "We need to give this government a chance.. They've been in 11 weeks and we need to allow them to make the changes.. To turn this oil tanker around.. Because it's going to take a number of years to get this right" "And if they don't do it, at the end of the day, they won't be in power i five years time"

Farrukh

386,459 просмотров • 1 год назад

So here is new local privilege escalation zero-day I discovered, not patched yet too :). In simple terms, if you have a service like RDP that exposes an RPC server, there many system services running as SYSTEM connect to it as RPC clients. If that service is turned off (RDP is off by default), it seems that any other process in Windows can expose the same RPC server using the same endpoint. Now all the RPC calls from that SYSTEM processes will come to this fake server and If the process that deployed the server has SeImpersonatePrivilege, it can escalate to SYSTEM by impersonate the RPC client. In the white paper below, I describe five exploit paths you can abuse. However it's architecture problem and maybe there are more. It's Not A Potato
1:49

So here is new local privilege escalation zero-day I discovered, not patched yet too :). In simple terms, if you have a service like RDP that exposes an RPC server, there many system services running as SYSTEM connect to it as RPC clients. If that service is turned off (RDP is off by default), it seems that any other process in Windows can expose the same RPC server using the same endpoint. Now all the RPC calls from that SYSTEM processes will come to this fake server and If the process that deployed the server has SeImpersonatePrivilege, it can escalate to SYSTEM by impersonate the RPC client. In the white paper below, I describe five exploit paths you can abuse. However it's architecture problem and maybe there are more. It's Not A Potato

Haidar

51,273 просмотров • 3 месяцев назад