Загрузка видео...

Не удалось загрузить видео

На главную

Offensive Security Tool: SecretOpt1c SecretOpt1c is written by Chris Abou-Chabké from Black Hat Ethical Hacking and its designed for Red Team, Pentesters, and Bug Bounty Hunters. It is a very powerful and versatile tool that helps uncover sensitive information on websites using Active and Passive Techniques for Superior Accuracy....

TheMsterDoctor1's profile picture

X

35,464 subscribers

20,768 просмотров • 1 год назад •via X (Twitter)

Наука и технологииНовости и политикаОбразование
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Offensive Security Tool: SQLMutant SQLMutant is written by Chris Abou-Chabké from Black Hat Ethical Hacking and its a comprehensive SQL injection tool. It uses a variety of techniques to detect vulnerabilities, including pattern matching, error analysis, and timing attacks, ensuring that no stone is left unturned. In addition to its powerful features, SQLMutant also integrates with Waybackurls and Arjun, enabling it to find even more parameters and merge all results together. This integration is highly beneficial as it provides automated subdomain enumeration and SQL injection testing capabilities. Requirements: To use SQLMutant, you need to have the following tools installed: figlet & lolcat: pip install lolcat & apt-get install figlet Waybackurls : go install Arjun: pip install arjun SQLMap: git clone --depth 1 sqlmap-dev Installation git clone cd SQLMutant chmod +x ./SQLMutant.sh Read the full post: #sqli #sqlinjection #hacking #bugbounty #pentesting #infosec #informationsecurity #offensivesecurity #redteam
0:15

Offensive Security Tool: SQLMutant SQLMutant is written by Chris Abou-Chabké from Black Hat Ethical Hacking and its a comprehensive SQL injection tool. It uses a variety of techniques to detect vulnerabilities, including pattern matching, error analysis, and timing attacks, ensuring that no stone is left unturned. In addition to its powerful features, SQLMutant also integrates with Waybackurls and Arjun, enabling it to find even more parameters and merge all results together. This integration is highly beneficial as it provides automated subdomain enumeration and SQL injection testing capabilities. Requirements: To use SQLMutant, you need to have the following tools installed: figlet & lolcat: pip install lolcat & apt-get install figlet Waybackurls : go install Arjun: pip install arjun SQLMap: git clone --depth 1 sqlmap-dev Installation git clone cd SQLMutant chmod +x ./SQLMutant.sh Read the full post: #sqli #sqlinjection #hacking #bugbounty #pentesting #infosec #informationsecurity #offensivesecurity #redteam

X

42,643 просмотров • 1 год назад

Offensive Security Tool: TerminatorZ TerminatorZ is developed by Chris Abou-Chabké from Black Hat Ethical Hacking and is an efficient web security tool designed to scan for potential vulnerabilities in your web applications. It uses a combination of advanced techniques and popular tools like 'waybackurls' and 'curl', to perform passive and quick scans, providing a rapid overview of potential issues. The results are displayed in an easy-to-read format in the terminal, with only the vulnerable findings saved for further investigation. 🟢New Version 2.0 Released! This update includes 8 NEW Proofs of Concept (POCs), bringing the total to 24 POCs, along with several other enhancements. ➡️New POCs: ◾File Upload Vulnerability ◾Command Injection ◾Host Header Injection ◾HTTP Parameter Pollution (HPP) ◾Clickjacking ◾CORS Misconfiguration ◾Sensitive Data Exposure ◾Session Fixation Read the post: Installation git clone cd TerminatorZ chmod +x ./TerminatorZ ⁠#securityassessments #pentesting #bugbounty #hacking #infosec #informationsecurity #cybersecurity #offensivesecurity #redteam
0:15

Offensive Security Tool: TerminatorZ TerminatorZ is developed by Chris Abou-Chabké from Black Hat Ethical Hacking and is an efficient web security tool designed to scan for potential vulnerabilities in your web applications. It uses a combination of advanced techniques and popular tools like 'waybackurls' and 'curl', to perform passive and quick scans, providing a rapid overview of potential issues. The results are displayed in an easy-to-read format in the terminal, with only the vulnerable findings saved for further investigation. 🟢New Version 2.0 Released! This update includes 8 NEW Proofs of Concept (POCs), bringing the total to 24 POCs, along with several other enhancements. ➡️New POCs: ◾File Upload Vulnerability ◾Command Injection ◾Host Header Injection ◾HTTP Parameter Pollution (HPP) ◾Clickjacking ◾CORS Misconfiguration ◾Sensitive Data Exposure ◾Session Fixation Read the post: Installation git clone cd TerminatorZ chmod +x ./TerminatorZ ⁠#securityassessments #pentesting #bugbounty #hacking #infosec #informationsecurity #cybersecurity #offensivesecurity #redteam

X

77,333 просмотров • 1 год назад

🛑 A powerful OSINT tool for exploring global sanctions data If you ever need to check sanctions across different authorities, this tool brings everything together in one place and it’s FREE to use. It’s called Sanctions Explorer, and it provides: ✅ Current + historical sanctions data 📚 ✅ Coverage from major bodies like the UN, EU, OFAC, and more ✅ Clean filtering to track individuals, companies, and entities ✅ A unified view instead of searching each authority separately 🌍 The project was initiated by Archer (a Berkeley-based nonprofit focused on human rights) and later fully developed by the C4ADS Data & Technology team known for their evidence-based research on global conflict and transnational security. If sanctions research is part of your OSINT workflow, this tool is worth bookmarking. And in the next post, I’ll share a quick OSINT tip for checking phone numbers on Telegram Check it out here: __________ P.S. ♻️ Repost if you found this helpful. If you liked this post and would like to learn more methods and techniques to discover information about people, check out my OSINT Mastery course
0:17

🛑 A powerful OSINT tool for exploring global sanctions data If you ever need to check sanctions across different authorities, this tool brings everything together in one place and it’s FREE to use. It’s called Sanctions Explorer, and it provides: ✅ Current + historical sanctions data 📚 ✅ Coverage from major bodies like the UN, EU, OFAC, and more ✅ Clean filtering to track individuals, companies, and entities ✅ A unified view instead of searching each authority separately 🌍 The project was initiated by Archer (a Berkeley-based nonprofit focused on human rights) and later fully developed by the C4ADS Data & Technology team known for their evidence-based research on global conflict and transnational security. If sanctions research is part of your OSINT workflow, this tool is worth bookmarking. And in the next post, I’ll share a quick OSINT tip for checking phone numbers on Telegram Check it out here: __________ P.S. ♻️ Repost if you found this helpful. If you liked this post and would like to learn more methods and techniques to discover information about people, check out my OSINT Mastery course

CyberSudo

28,153 просмотров • 9 месяцев назад

What makes people stop scrolling on X? What drives engagement? What triggers the algorithm? For the past year, we’ve reverse-engineered hundreds of viral tweets from my account and scientifically studied the X algorithm. We embedded every insight we gained into SurgeKit. It is a powerful tool that helps you write better on X, post smarter, and grow faster. Used by top creators. Built by real practitioners. Fed with virality data from my own content. Now in public beta →
0:10

What makes people stop scrolling on X? What drives engagement? What triggers the algorithm? For the past year, we’ve reverse-engineered hundreds of viral tweets from my account and scientifically studied the X algorithm. We embedded every insight we gained into SurgeKit. It is a powerful tool that helps you write better on X, post smarter, and grow faster. Used by top creators. Built by real practitioners. Fed with virality data from my own content. Now in public beta →

Mario Nawfal

242,681 просмотров • 1 год назад

Karpathy's Agentic Engineering finally has proper tooling! (built by Google) Karpathy defined agentic engineering as the discipline that separates production agent work from vibe coding. The core skills he listed were spec design, eval loops, and security oversight. The problem has been that practicing this still requires a different tool for every phase: - editor for code - a terminal for scaffolding - a browser for testing - a cloud console for deployment - and a separate framework for evals. Every transition is a context switch. The solution to production-grade Agentic Engineering is now actually implemented in Google’s Agents CLI. It covers the entire workflow in one place for scaffolding, evaluating, and deploying ADK agents. One setup command injects 7 ADK-specific skills into a coding agent's context, which lets it handle scaffolding, evals, deployment, and enterprise registration through natural language. I tested this end-to-end by building a RAG agent from scratch using Claude Code. It scaffolded the full project from the ADK agentic_rag template, generated 20 eval scenarios with LLM-as-judge scoring, and returned a quantitative scorecard. Finally, it also deployed everything to Agent Runtime and registered the agent to Gemini Enterprise, so the entire org can discover and use it. The video below shows this in action, and I worked with the Google Cloud team to put this together. Agents CLI GitHub repo → (don't forget to star it ⭐ ) I wrote up the full build covering all six steps from install to enterprise registration. It includes the eval scorecard, the instruction loophole the eval caught before deployment, and what the deployment process actually looks like end-to-end. Read it below.
0:15

Karpathy's Agentic Engineering finally has proper tooling! (built by Google) Karpathy defined agentic engineering as the discipline that separates production agent work from vibe coding. The core skills he listed were spec design, eval loops, and security oversight. The problem has been that practicing this still requires a different tool for every phase: - editor for code - a terminal for scaffolding - a browser for testing - a cloud console for deployment - and a separate framework for evals. Every transition is a context switch. The solution to production-grade Agentic Engineering is now actually implemented in Google’s Agents CLI. It covers the entire workflow in one place for scaffolding, evaluating, and deploying ADK agents. One setup command injects 7 ADK-specific skills into a coding agent's context, which lets it handle scaffolding, evals, deployment, and enterprise registration through natural language. I tested this end-to-end by building a RAG agent from scratch using Claude Code. It scaffolded the full project from the ADK agentic_rag template, generated 20 eval scenarios with LLM-as-judge scoring, and returned a quantitative scorecard. Finally, it also deployed everything to Agent Runtime and registered the agent to Gemini Enterprise, so the entire org can discover and use it. The video below shows this in action, and I worked with the Google Cloud team to put this together. Agents CLI GitHub repo → (don't forget to star it ⭐ ) I wrote up the full build covering all six steps from install to enterprise registration. It includes the eval scorecard, the instruction loophole the eval caught before deployment, and what the deployment process actually looks like end-to-end. Read it below.

Akshay 🚀

258,823 просмотров • 3 месяцев назад

Boom! Grok Tasks Make It One Of The Most POWERFUL Real-Time AI Systems In The World. — My How to Use Grok Tasks With Hidden Tools For Powerful Daily Output. Grok Tasks are customizable AI workflows that integrate a variety of tools to streamline daily activities, from research and analysis to creative planning and problem-solving. I have been using them for quite sometime and because of the vital heartbeat of news and first person data on X, it is the most powerful AI platform available. By combining Tasks with tools like web searches, X platform interactions, code execution, and media viewers, you can build efficient, automated processes. These tasks work by prompting Grok with a clear description of what you want to achieve, and Grok will intelligently call the necessary tools in sequence or parallel to deliver results. Here's a step-by-step guide to creating and using Grok Tasks: Step 1: Define Your Task Start by clearly outlining the daily activity or goal. Consider what inputs you have (e.g., a URL, a query, or an attachment) and what output you need (e.g., a summary, calculation, or visual analysis). Break it down into subtasks to identify tool needs. For example, if your task involves researching current events, note that you'll need search and browsing capabilities. Step 2: Review Available Tools Familiarize yourself with the tools Grok can access. Here's a quick overview: - Code Execution: Run Python code for calculations, data processing, or simulations using libraries like numpy, pandas, or sympy. - Browse Page: Fetch and summarize content from any website URL with custom instructions. - Web Search: Perform general internet searches, returning results with optional operators like site:. - Web Search With Snippets: Get quick, detailed excerpts from search results for fact-checking. - X Keyword Search: Advanced search for X posts using operators like from:, since:, or filter:. - X Semantic Search: Find semantically related X posts based on a query, with filters for dates or users. - X User Search: Locate X users by name or handle. - X Thread Fetch: Retrieve a full X post thread, including context like replies and parents. - View Image: Analyze an image from a URL or conversation ID. - View X Video: Extract frames and subtitles from an X-hosted video. - Search PDF Attachment: Query a PDF file for relevant pages using keyword or regex modes. - Browse PDF Attachment: View specific pages of a PDF with text and screenshots. Select tools that align with your task. Aim for a mix to handle data gathering, processing, and visualization. Step 3: Craft Your Prompt Write a detailed prompt to Grok describing the task. Include: - The overall goal. - Specific steps or subtasks. - References to tools if you want to guide the process (e.g., "Use web_search to find sources, then code_execution to analyze data"). - Any constraints, like dates or limits. Example prompt: "Create a Grok Task for my morning routine: Search recent X posts about tech news using x_keyword_search, fetch a key thread with x_thread_fetch, and summarize with browse_page on linked articles." Step 4: Submit and Interact Send your prompt to Grok. It will process the task by calling tools as needed, often in parallel for efficiency. Review the output and refine with follow-up prompts if required (e.g., "Expand on that using view_image for visuals"). Iterate to fine-tune the workflow for reuse. Step 5: Save and Reuse Once refined, note the prompt as a template for future use. You can adapt it for similar tasks, making Grok Tasks a habitual part of your day. Finding Grok Tasks To discover existing Grok Tasks or inspiration for new ones, use X searches with tools like x_keyword_search or x_semantic_search (e.g., query: "Grok Tasks examples" with mode: Latest). Browse community-shared threads via x_thread_fetch, or web_search for tutorials on xAI features. Prompt Grok directly: "Show me popular Grok Tasks for productivity." 1 of 3
0:12

Boom! Grok Tasks Make It One Of The Most POWERFUL Real-Time AI Systems In The World. — My How to Use Grok Tasks With Hidden Tools For Powerful Daily Output. Grok Tasks are customizable AI workflows that integrate a variety of tools to streamline daily activities, from research and analysis to creative planning and problem-solving. I have been using them for quite sometime and because of the vital heartbeat of news and first person data on X, it is the most powerful AI platform available. By combining Tasks with tools like web searches, X platform interactions, code execution, and media viewers, you can build efficient, automated processes. These tasks work by prompting Grok with a clear description of what you want to achieve, and Grok will intelligently call the necessary tools in sequence or parallel to deliver results. Here's a step-by-step guide to creating and using Grok Tasks: Step 1: Define Your Task Start by clearly outlining the daily activity or goal. Consider what inputs you have (e.g., a URL, a query, or an attachment) and what output you need (e.g., a summary, calculation, or visual analysis). Break it down into subtasks to identify tool needs. For example, if your task involves researching current events, note that you'll need search and browsing capabilities. Step 2: Review Available Tools Familiarize yourself with the tools Grok can access. Here's a quick overview: - Code Execution: Run Python code for calculations, data processing, or simulations using libraries like numpy, pandas, or sympy. - Browse Page: Fetch and summarize content from any website URL with custom instructions. - Web Search: Perform general internet searches, returning results with optional operators like site:. - Web Search With Snippets: Get quick, detailed excerpts from search results for fact-checking. - X Keyword Search: Advanced search for X posts using operators like from:, since:, or filter:. - X Semantic Search: Find semantically related X posts based on a query, with filters for dates or users. - X User Search: Locate X users by name or handle. - X Thread Fetch: Retrieve a full X post thread, including context like replies and parents. - View Image: Analyze an image from a URL or conversation ID. - View X Video: Extract frames and subtitles from an X-hosted video. - Search PDF Attachment: Query a PDF file for relevant pages using keyword or regex modes. - Browse PDF Attachment: View specific pages of a PDF with text and screenshots. Select tools that align with your task. Aim for a mix to handle data gathering, processing, and visualization. Step 3: Craft Your Prompt Write a detailed prompt to Grok describing the task. Include: - The overall goal. - Specific steps or subtasks. - References to tools if you want to guide the process (e.g., "Use web_search to find sources, then code_execution to analyze data"). - Any constraints, like dates or limits. Example prompt: "Create a Grok Task for my morning routine: Search recent X posts about tech news using x_keyword_search, fetch a key thread with x_thread_fetch, and summarize with browse_page on linked articles." Step 4: Submit and Interact Send your prompt to Grok. It will process the task by calling tools as needed, often in parallel for efficiency. Review the output and refine with follow-up prompts if required (e.g., "Expand on that using view_image for visuals"). Iterate to fine-tune the workflow for reuse. Step 5: Save and Reuse Once refined, note the prompt as a template for future use. You can adapt it for similar tasks, making Grok Tasks a habitual part of your day. Finding Grok Tasks To discover existing Grok Tasks or inspiration for new ones, use X searches with tools like x_keyword_search or x_semantic_search (e.g., query: "Grok Tasks examples" with mode: Latest). Browse community-shared threads via x_thread_fetch, or web_search for tutorials on xAI features. Prompt Grok directly: "Show me popular Grok Tasks for productivity." 1 of 3

Brian Roemmele

152,242 просмотров • 9 месяцев назад

BREAKING: Anthropic just dropped Opus 4.8—and it is a MONSTER We've been testing for about a week Every 📧 and our verdict is they could've just called it Opus 5, it's that good. Here's our vibe check: - Beats GPT-5.5 on Senior Engineer bench. On our toughest benchmark Opus 4.8 scores a 63—a hair higher than GPT-5.5's score of 62, and a full 30 points higher than Opus 4.7. It tackled a ground-up rewrite of a production codebase, and actually built something that works. HOWEVER: Coding performance varied a lot at different reasoning levels. We recommend using it on xhigh for best results. - Incredibly good writer. Opus 4.8 scored a 79.6 on our writing benchmark—measuring models on real-world writing tasks we do all of the time like essay writing, promo email writing, and more. It beats GPT-5.5 by 6 points. It produces well-written prose with fewer "AI-isms". It's also very good at writing in your voice given the right context. HOWEVER: Writing performance also varied with reasoning levels. Medium reasoning had higher incidence of AI-isms—we found best results with high. - Beast at knowledge work. Opus 4.8 is very good at general knowledge work tasks like report creation, research and more. It produced the best PowerPoint one-shot we've ever seen on our deck generation benchmark. - Emotionally intelligent, willing to question the frame. I've also found it to be quite good at talking through psychological or interpersonal issues. It has a high EQ, and it's also good at not glazing and helping to expand your perspective. Its thought process feels extremely rich and dynamic. THE BAD: These days a model is only as good as its harness, and Codex is still a far superior harness to the Claude Desktop app. This has kept me using Codex + GPT-5.5 as my daily driver, but I am flipping back and forth a lot more between Codex and Claude. Anthropic is back baby! Read the rest on Every 📧:
0:11

BREAKING: Anthropic just dropped Opus 4.8—and it is a MONSTER We've been testing for about a week Every 📧 and our verdict is they could've just called it Opus 5, it's that good. Here's our vibe check: - Beats GPT-5.5 on Senior Engineer bench. On our toughest benchmark Opus 4.8 scores a 63—a hair higher than GPT-5.5's score of 62, and a full 30 points higher than Opus 4.7. It tackled a ground-up rewrite of a production codebase, and actually built something that works. HOWEVER: Coding performance varied a lot at different reasoning levels. We recommend using it on xhigh for best results. - Incredibly good writer. Opus 4.8 scored a 79.6 on our writing benchmark—measuring models on real-world writing tasks we do all of the time like essay writing, promo email writing, and more. It beats GPT-5.5 by 6 points. It produces well-written prose with fewer "AI-isms". It's also very good at writing in your voice given the right context. HOWEVER: Writing performance also varied with reasoning levels. Medium reasoning had higher incidence of AI-isms—we found best results with high. - Beast at knowledge work. Opus 4.8 is very good at general knowledge work tasks like report creation, research and more. It produced the best PowerPoint one-shot we've ever seen on our deck generation benchmark. - Emotionally intelligent, willing to question the frame. I've also found it to be quite good at talking through psychological or interpersonal issues. It has a high EQ, and it's also good at not glazing and helping to expand your perspective. Its thought process feels extremely rich and dynamic. THE BAD: These days a model is only as good as its harness, and Codex is still a far superior harness to the Claude Desktop app. This has kept me using Codex + GPT-5.5 as my daily driver, but I am flipping back and forth a lot more between Codex and Claude. Anthropic is back baby! Read the rest on Every 📧:

Dan Shipper

354,673 просмотров • 4 месяцев назад

Announcing Centaur 2.0! Centaur is frontier, agentic infrastructure that you own. Centaur is like Claude Tag, but open source and on steroids. Centaur 2.0 can connect to everything you have access to, and can reason over it next to where you work, either in Slack, Discord, Teams or in your local Codex or Claude Code via an MCP. But context isn't useful if it can be accessed by anyone, so we rebuilt Centaur from the ground up for security. Centaur obviously doesn't have access to secrets because we're leveraging egress proxies. Centaur 2.0 takes that a step further enabling administrators to configure who can access what from where. This means that Centaur can have access to private information that you normally wouldn't feel comfortable giving it access to (e.g. DMs or sensitive channels and docs) but only expose it to authorized principals. It also means you can add Centaur to external channels and leverage it as a virtual colleague that doesn't live only in your Slack, for example, but also your Slack Connect channels! This is extremely powerful as we start moving to a world where agents cross organizational boundaries. Also in case anyone's wondering, yes we rewrote it in Rust! Centaur is now way way more stable at durably executing threads, and its workflow engine is now based on Absurd. We have been operating Centaur since January, officially launched and Open Sourced it in May, and now we're full speed towards making it the #1 open source agentic infrastructure. To succeed at that, I'm thrilled to welcome Matthew Slipper to the Paradigm team who will be leading all our Applied AI work, while continuing to maintain and extend Iron Proxy as the leading secure secret access for agents. Welcome Matt, it's an honor after all these years of knowing you! Read the full blogpost below, and apply to join our team!
0:19

Announcing Centaur 2.0! Centaur is frontier, agentic infrastructure that you own. Centaur is like Claude Tag, but open source and on steroids. Centaur 2.0 can connect to everything you have access to, and can reason over it next to where you work, either in Slack, Discord, Teams or in your local Codex or Claude Code via an MCP. But context isn't useful if it can be accessed by anyone, so we rebuilt Centaur from the ground up for security. Centaur obviously doesn't have access to secrets because we're leveraging egress proxies. Centaur 2.0 takes that a step further enabling administrators to configure who can access what from where. This means that Centaur can have access to private information that you normally wouldn't feel comfortable giving it access to (e.g. DMs or sensitive channels and docs) but only expose it to authorized principals. It also means you can add Centaur to external channels and leverage it as a virtual colleague that doesn't live only in your Slack, for example, but also your Slack Connect channels! This is extremely powerful as we start moving to a world where agents cross organizational boundaries. Also in case anyone's wondering, yes we rewrote it in Rust! Centaur is now way way more stable at durably executing threads, and its workflow engine is now based on Absurd. We have been operating Centaur since January, officially launched and Open Sourced it in May, and now we're full speed towards making it the #1 open source agentic infrastructure. To succeed at that, I'm thrilled to welcome Matthew Slipper to the Paradigm team who will be leading all our Applied AI work, while continuing to maintain and extend Iron Proxy as the leading secure secret access for agents. Welcome Matt, it's an honor after all these years of knowing you! Read the full blogpost below, and apply to join our team!

Georgios Konstantopoulos

144,862 просмотров • 1 месяц назад

In 2025, the AgentFlayer exploit highlighted a new category of risk in AI systems. It was not a traditional breach involving stolen credentials or broken encryption. Instead, it demonstrated how an autonomous AI agent could be manipulated into executing unintended actions by processing malicious instructions embedded inside content it automatically processes. The incident did not expose a flaw in one specific integration. It revealed a structural weakness in how many modern AI agents are built. Today’s agents are no longer passive language models. They read documents automatically, scan emails, connect to SaaS tools, access cloud storage, and execute actions across multiple systems. To be useful, they are granted meaningful permissions. That capability creates value, but it also expands the attack surface. Most agent environments operate in a trusted, plaintext execution model. Data is encrypted at rest and in transit, but it is typically decrypted during inference so the model can process it. That runtime visibility is where potential risk lies. In a zero-click scenario like AgentFlayer, an attacker can embed hidden instructions inside a document that the AI processes automatically. Because the agent may have access to connected systems such as Google Drive, Slack, or GitHub, it can potentially be influenced to retrieve sensitive information or perform unintended actions. The user does not need to click a malicious link or approve a suspicious request. Therefore, the core issue is that during execution, the system may have access to sensitive data and broad privileges, meaning whoever controls the execution environment ultimately controls access to that data. Now consider a different architectural approach. If a system is designed so that data remains protected during execution, the risk profile changes. On Nesa, privacy is enforced at the execution layer through Equivariant Encryption. Computation can occur on encrypted data, reducing the visibility surface during runtime. Sensitive inputs and models do not need to be exposed in plain text to infrastructure operators for inference to occur. This does not eliminate prompt injection, logic manipulation, or tool misuse. Encryption alone cannot prevent an agent from being instructed to take an unintended action if it has been granted that permission. What it does do is materially reduce confidentiality risk. By limiting access to readable sensitive data during execution and reducing unilateral visibility at the infrastructure layer, the potential blast radius of a successful manipulation attempt is constrained. As AI agents become more autonomous and embedded into enterprise workflows, security must move deeper into architecture. The goal is not to claim invulnerability. It is to reduce trust concentration and contain systemic exposure when failures occur. AgentFlayer was not simply a one-off exploit. It was a reminder that in autonomous systems, execution-layer design determines how risk propagates.
0:11

In 2025, the AgentFlayer exploit highlighted a new category of risk in AI systems. It was not a traditional breach involving stolen credentials or broken encryption. Instead, it demonstrated how an autonomous AI agent could be manipulated into executing unintended actions by processing malicious instructions embedded inside content it automatically processes. The incident did not expose a flaw in one specific integration. It revealed a structural weakness in how many modern AI agents are built. Today’s agents are no longer passive language models. They read documents automatically, scan emails, connect to SaaS tools, access cloud storage, and execute actions across multiple systems. To be useful, they are granted meaningful permissions. That capability creates value, but it also expands the attack surface. Most agent environments operate in a trusted, plaintext execution model. Data is encrypted at rest and in transit, but it is typically decrypted during inference so the model can process it. That runtime visibility is where potential risk lies. In a zero-click scenario like AgentFlayer, an attacker can embed hidden instructions inside a document that the AI processes automatically. Because the agent may have access to connected systems such as Google Drive, Slack, or GitHub, it can potentially be influenced to retrieve sensitive information or perform unintended actions. The user does not need to click a malicious link or approve a suspicious request. Therefore, the core issue is that during execution, the system may have access to sensitive data and broad privileges, meaning whoever controls the execution environment ultimately controls access to that data. Now consider a different architectural approach. If a system is designed so that data remains protected during execution, the risk profile changes. On Nesa, privacy is enforced at the execution layer through Equivariant Encryption. Computation can occur on encrypted data, reducing the visibility surface during runtime. Sensitive inputs and models do not need to be exposed in plain text to infrastructure operators for inference to occur. This does not eliminate prompt injection, logic manipulation, or tool misuse. Encryption alone cannot prevent an agent from being instructed to take an unintended action if it has been granted that permission. What it does do is materially reduce confidentiality risk. By limiting access to readable sensitive data during execution and reducing unilateral visibility at the infrastructure layer, the potential blast radius of a successful manipulation attempt is constrained. As AI agents become more autonomous and embedded into enterprise workflows, security must move deeper into architecture. The goal is not to claim invulnerability. It is to reduce trust concentration and contain systemic exposure when failures occur. AgentFlayer was not simply a one-off exploit. It was a reminder that in autonomous systems, execution-layer design determines how risk propagates.

Nesa

17,038 просмотров • 7 месяцев назад

The agents that run my LinkedIn outreach are currently getting a 55% reply rate on cold outreach and are generating hundreds of signups for me completely on autopilot. I packaged the exact system into a skill that works for any harness so you can copy it for yourself. Heres why this system prints qualified replies: most outbound starts with a list out of apollo, zoominfo or clay. firmographics. who could buy. that list is cold on arrival because "could buy" and "is thinking about this right now" are two different lists. so mine starts from linkedin posts my buyers are already reacting to. that also means every person on the list actually uses linkedin, since they just reacted to a post on it. most people don't, so a message to a normal list mostly sits unread. across the client accounts i've worked on, an apollo-style list usually gets about 1/10th the reply rate of one built this way. the posts come from three places: buyers posting about the problem themselves (hiring for it, breaking under it, shopping for a fix), a watchlist of my competitors and the creators my buyers follow, and the topic itself. when a buyer writes the post, the author is the best lead on it. anyone reacting to a competitor's product post is already shopping. every search is written in the words buyers actually post in, because vendor jargon only finds other vendors. it pulls everyone who reacted or commented, drops anyone i've already contacted plus competitors and their staff, and scores each person 0 to 100 against my ICP with the reason written on the row. anyone under the bar is gone before i've paid for their email. this is the part everyone skips. for every lead that clears the bar, it reads their website, their LinkedIn profile & anything else it can find about them on the internet, works out what agent would actually help their operation, and builds it. a real one, sitting there, with a claim link. so i am not sending "quick question about your outbound." i am sending "i built you an agent that does X for your company, want me to send over the link?" it sequences that into heyreach for linkedin and instantly for email, then runs the inbox. anyone who has done outbound knows reply speed is most of the game, and also the exact thing you quit doing by week three. every reply fires a webhook. it re-reads the whole thread, tags it interested, not interested or auto-reply, pings me in slack the second someone's interested, and sends them their claim link. i don't touch it. and every reply feeds the next run. it writes more searches like the ones that got answers, so week four sources better than week one. the general rule, and it has nothing to do with my software: use AI to build a personalized, genuinely useful asset and give it away in the first message. you're not hard selling anyone. you are handing them a thing. and because it's built off their actual company, nobody can mistake it for a template. reply rates go up for both reasons at once. it doesn't have to be an agent. the skill builds whatever your business is already good at making: - sell lead gen? a sample list of 10 to 25 of their own buyers - sell SEO? the 3 to 5 searches their competitors win and they don't - build websites? a teardown of their homepage with three fixes - consult? a one-page plan for the exact problem they commented about the playbook behind the messages is the best outbound advice on the internet in one place. it's built from 25 full-length videos and 60 posts from operators who do this at scale, plus 3 production systems running this exact play. scraped the videos and posts with superagnt's data MCP btw. run it once in claude code or codex and look at what it staged. once you like it, it turns the whole thing into a small agent team on a schedule. the sourcing side runs about 8 cents per qualified lead in production. grab it here, paste the install prompt into your agent and let it cook:
0:14

The agents that run my LinkedIn outreach are currently getting a 55% reply rate on cold outreach and are generating hundreds of signups for me completely on autopilot. I packaged the exact system into a skill that works for any harness so you can copy it for yourself. Heres why this system prints qualified replies: most outbound starts with a list out of apollo, zoominfo or clay. firmographics. who could buy. that list is cold on arrival because "could buy" and "is thinking about this right now" are two different lists. so mine starts from linkedin posts my buyers are already reacting to. that also means every person on the list actually uses linkedin, since they just reacted to a post on it. most people don't, so a message to a normal list mostly sits unread. across the client accounts i've worked on, an apollo-style list usually gets about 1/10th the reply rate of one built this way. the posts come from three places: buyers posting about the problem themselves (hiring for it, breaking under it, shopping for a fix), a watchlist of my competitors and the creators my buyers follow, and the topic itself. when a buyer writes the post, the author is the best lead on it. anyone reacting to a competitor's product post is already shopping. every search is written in the words buyers actually post in, because vendor jargon only finds other vendors. it pulls everyone who reacted or commented, drops anyone i've already contacted plus competitors and their staff, and scores each person 0 to 100 against my ICP with the reason written on the row. anyone under the bar is gone before i've paid for their email. this is the part everyone skips. for every lead that clears the bar, it reads their website, their LinkedIn profile & anything else it can find about them on the internet, works out what agent would actually help their operation, and builds it. a real one, sitting there, with a claim link. so i am not sending "quick question about your outbound." i am sending "i built you an agent that does X for your company, want me to send over the link?" it sequences that into heyreach for linkedin and instantly for email, then runs the inbox. anyone who has done outbound knows reply speed is most of the game, and also the exact thing you quit doing by week three. every reply fires a webhook. it re-reads the whole thread, tags it interested, not interested or auto-reply, pings me in slack the second someone's interested, and sends them their claim link. i don't touch it. and every reply feeds the next run. it writes more searches like the ones that got answers, so week four sources better than week one. the general rule, and it has nothing to do with my software: use AI to build a personalized, genuinely useful asset and give it away in the first message. you're not hard selling anyone. you are handing them a thing. and because it's built off their actual company, nobody can mistake it for a template. reply rates go up for both reasons at once. it doesn't have to be an agent. the skill builds whatever your business is already good at making: - sell lead gen? a sample list of 10 to 25 of their own buyers - sell SEO? the 3 to 5 searches their competitors win and they don't - build websites? a teardown of their homepage with three fixes - consult? a one-page plan for the exact problem they commented about the playbook behind the messages is the best outbound advice on the internet in one place. it's built from 25 full-length videos and 60 posts from operators who do this at scale, plus 3 production systems running this exact play. scraped the videos and posts with superagnt's data MCP btw. run it once in claude code or codex and look at what it staged. once you like it, it turns the whole thing into a small agent team on a schedule. the sourcing side runs about 8 cents per qualified lead in production. grab it here, paste the install prompt into your agent and let it cook:

Jáen

295,995 просмотров • 9 дней назад

InterLink’s Early Vision for NIST-Standardised Post-Quantum Cryptography 🔐✨ The next five years may bring a much clearer answer to a question that has long been difficult to judge: Is a digital asset truly secure? 🤔 For InterLink, the answer may increasingly depend on one critical factor: whether it is quantum-resistant and aligned with NIST standards 🧬🔒 Why this matters now ⚠️ Two fast-moving technologies are reshaping digital security: • AI is improving the ability to discover weaknesses in systems 🤖 • Quantum computing is advancing towards the point where today’s cryptographic foundations could become vulnerable ⚛️ For InterLink, this is not just a theoretical discussion. It is a reminder that blockchain networks, wallets, and custody systems must prepare now for the post-quantum era ⏳🛡️ Why NIST is central to InterLink’s approach 📘🏛️ Many in the InterLink community already know NIST. NIST, part of the U.S. Department of Commerce, plays a major role in defining and evaluating security standards, including those for post-quantum cryptography. Its work matters because it helps shape what “secure” will mean in a quantum-capable future 📊🔍 In practical terms, InterLink’s long-term security vision is closely tied to whether its cryptographic design can withstand post-quantum threats 🚀 The risk for blockchain networks and digital assets 🔎💥 Recent research and experiments from major organisations, including Google, have highlighted a growing concern: what was once considered extremely difficult, using quantum computing to threaten cryptographic systems, is no longer something that can be ignored 🧪⚠️ That does not mean current systems are broken today. It does mean that networks which fail to prepare for post-quantum threats could face serious risks later 📉 For InterLink, this is exactly why early research matters 🧠✨ If sufficiently powerful quantum computers become available, some current cryptographic methods may become vulnerable. For any network storing value, identity, NFTs, or permissions, that is a major issue 💳🖼️🧾 InterLink’s early work on post-quantum readiness 🧠🔐 InterLink Foundation has already been researching: ✅ digital signatures ✍️ ✅ cryptographic algorithms 🔣 ✅ migration mechanisms for future security upgrades 🔄 One of the most notable areas of work is the ability to generate new private keys from an existing seed phrase 🌱➡️🔑 This matters because it offers a pathway to improve security without forcing users to abandon access to their assets 🙌 Address Alias: preserving continuity during migration 🪪🔗 Another important InterLink mechanism is Address Alias. This is designed to let users: ✓ retain their existing wallet addresses 🧾 ✓ preserve associated tokens and NFTs 🖼️💰 ✓ migrate to a new cryptographic security architecture 🔐➡️🛠️ That is a practical and user-friendly design choice. Security upgrades are often hard to adopt when they break continuity. InterLink’s approach aims to solve that problem 🌉 Bringing post-quantum protection into smart contracts 🛡️📜 InterLink is also implementing SLH-DSA-SHA2-128s (FIPS 205) within IRC smart contracts. This adds another layer of protection for: • vaults 🏦 • high-value assets 💎 • long-term storage 📦 • sensitive on-chain operations ⚙️ The goal is not only to protect wallets, but also to strengthen the systems that govern custody and transaction security across the network 🧱🔒 Testing on the Taj Mahal Testnet 🧪🛰️ These experiments are currently being conducted on the InterLink Taj Mahal Testnet. According to InterLink, the experimental implementations have passed the NIST-based simulation tests carried out so far ✅📈 That is an encouraging early signal, although broader testing and real-world validation will remain important as development continues 🔍 Looking ahead to 2027 🚀🌍 InterLink’s stated goal is to fully integrate this architecture into the Open Mainnet in 2027. If achieved, that would bring InterLink closer to a future where security is defined not only by current best practice, but by resilience against quantum-era threats 🛡️⚛️ The bigger takeaway 🌍✨ The key lesson is simple: In the quantum era, security will not only mean protecting your private will mean asking whether the cryptography behind that key was built to survive the next generation of computing 🔐⏭️ For InterLink, this is a strategic direction with long-term significance 📌 Final thought 💡 Post-quantum readiness is no longer just a technical topic for specialists. For InterLink, it is becoming part of the broader conversation about long-term digital asset security 🧠🔒 NIST-aligned cryptography, practical migration paths, and user-preserving design may soon define the networks people trust most 🌟 InterLink Labs 👤 + 🌐 KV Reina | InterLink Labs InterLink Foundation #InterLink #ITLG #ITL #WeAreTheFirst10MLinkers Join me on InterLink 😁 Start mining now and use my invitation link: 💰 My code is: 111222777888 💰 Please DM me once you have used my code. 👍
0:11

InterLink’s Early Vision for NIST-Standardised Post-Quantum Cryptography 🔐✨ The next five years may bring a much clearer answer to a question that has long been difficult to judge: Is a digital asset truly secure? 🤔 For InterLink, the answer may increasingly depend on one critical factor: whether it is quantum-resistant and aligned with NIST standards 🧬🔒 Why this matters now ⚠️ Two fast-moving technologies are reshaping digital security: • AI is improving the ability to discover weaknesses in systems 🤖 • Quantum computing is advancing towards the point where today’s cryptographic foundations could become vulnerable ⚛️ For InterLink, this is not just a theoretical discussion. It is a reminder that blockchain networks, wallets, and custody systems must prepare now for the post-quantum era ⏳🛡️ Why NIST is central to InterLink’s approach 📘🏛️ Many in the InterLink community already know NIST. NIST, part of the U.S. Department of Commerce, plays a major role in defining and evaluating security standards, including those for post-quantum cryptography. Its work matters because it helps shape what “secure” will mean in a quantum-capable future 📊🔍 In practical terms, InterLink’s long-term security vision is closely tied to whether its cryptographic design can withstand post-quantum threats 🚀 The risk for blockchain networks and digital assets 🔎💥 Recent research and experiments from major organisations, including Google, have highlighted a growing concern: what was once considered extremely difficult, using quantum computing to threaten cryptographic systems, is no longer something that can be ignored 🧪⚠️ That does not mean current systems are broken today. It does mean that networks which fail to prepare for post-quantum threats could face serious risks later 📉 For InterLink, this is exactly why early research matters 🧠✨ If sufficiently powerful quantum computers become available, some current cryptographic methods may become vulnerable. For any network storing value, identity, NFTs, or permissions, that is a major issue 💳🖼️🧾 InterLink’s early work on post-quantum readiness 🧠🔐 InterLink Foundation has already been researching: ✅ digital signatures ✍️ ✅ cryptographic algorithms 🔣 ✅ migration mechanisms for future security upgrades 🔄 One of the most notable areas of work is the ability to generate new private keys from an existing seed phrase 🌱➡️🔑 This matters because it offers a pathway to improve security without forcing users to abandon access to their assets 🙌 Address Alias: preserving continuity during migration 🪪🔗 Another important InterLink mechanism is Address Alias. This is designed to let users: ✓ retain their existing wallet addresses 🧾 ✓ preserve associated tokens and NFTs 🖼️💰 ✓ migrate to a new cryptographic security architecture 🔐➡️🛠️ That is a practical and user-friendly design choice. Security upgrades are often hard to adopt when they break continuity. InterLink’s approach aims to solve that problem 🌉 Bringing post-quantum protection into smart contracts 🛡️📜 InterLink is also implementing SLH-DSA-SHA2-128s (FIPS 205) within IRC smart contracts. This adds another layer of protection for: • vaults 🏦 • high-value assets 💎 • long-term storage 📦 • sensitive on-chain operations ⚙️ The goal is not only to protect wallets, but also to strengthen the systems that govern custody and transaction security across the network 🧱🔒 Testing on the Taj Mahal Testnet 🧪🛰️ These experiments are currently being conducted on the InterLink Taj Mahal Testnet. According to InterLink, the experimental implementations have passed the NIST-based simulation tests carried out so far ✅📈 That is an encouraging early signal, although broader testing and real-world validation will remain important as development continues 🔍 Looking ahead to 2027 🚀🌍 InterLink’s stated goal is to fully integrate this architecture into the Open Mainnet in 2027. If achieved, that would bring InterLink closer to a future where security is defined not only by current best practice, but by resilience against quantum-era threats 🛡️⚛️ The bigger takeaway 🌍✨ The key lesson is simple: In the quantum era, security will not only mean protecting your private will mean asking whether the cryptography behind that key was built to survive the next generation of computing 🔐⏭️ For InterLink, this is a strategic direction with long-term significance 📌 Final thought 💡 Post-quantum readiness is no longer just a technical topic for specialists. For InterLink, it is becoming part of the broader conversation about long-term digital asset security 🧠🔒 NIST-aligned cryptography, practical migration paths, and user-preserving design may soon define the networks people trust most 🌟 InterLink Labs 👤 + 🌐 KV Reina | InterLink Labs InterLink Foundation #InterLink #ITLG #ITL #WeAreTheFirst10MLinkers Join me on InterLink 😁 Start mining now and use my invitation link: 💰 My code is: 111222777888 💰 Please DM me once you have used my code. 👍

Tekkaus® | InterLink • MOD • T2 Community Builder

22,550 просмотров • 1 месяц назад

how consumer apps are making money from tiktok slideshows... most people think app marketing means paid ads or influencer deals. the apps actually printing money right now are doing neither. they're running networks of slideshow accounts that look like normal content pages, and almost nobody notices the machine behind them. here's how it works. you follow what seems like a regular relationship advice account. the posts are relatable, sometimes rude, exactly what the algorithm feeds you. then you hit slide 5 and there's an app mentioned. every single post on that account funnels to the same app. one publisher running this playbook did $40k last month, verified by sensor tower, and that's one publisher. the part people miss: most of these accounts aren't even run by humans making content. the characters are AI. the images are AI. the "person" whose thoughts you're reading doesn't exist. viewers aren't fact-checking, they're reacting to whether the post is relatable, and AI passes that bar daily. myths keeping people out of this: "i need to make original content for my app." false. the fastest path is taking a slideshow that's already proven to convert, keeping the structure and pacing exactly, and swapping only the ad slide. the performance lives in the format, not in your creativity. "i can find my images on google." false. google image search is flooded with stock assets that read as fake instantly. pinterest is where the authentic-looking references live. and you don't post those directly, they're copyrighted, you feed them to AI as reference and generate your own singular images. "every post needs to go viral." false. accounts in this niche post variations constantly and most land under a thousand views. that's the model working. the flopped posts still collect saves and likes, and one breakout carries the month. volume across variations beats one polished post every time. "the ad slide should change with each post." false. everything around it should vary, the pain point, the images, the captions. the ad slide stays identical and polished every single time. you write that caption once, make it perfect, and lock it. consistency on the conversion slide, chaos everywhere else. what actually drives this: you can now clone a proven slideshow into a hundred variations, schedule a month of posts in advance, and have an ai agent handle the posting from your terminal. the production bottleneck is gone. the only real constraint left is picking a post that's already validated and moving before the account farms make this the default playbook for every app on the store. structural notes on what i changed: the original was a chronological tutorial, this reframes it as a reveal (the machine exists, here's how it works) with the tool mechanics compressed into the final paragraph instead of a walkthrough. the myths section absorbs the google/pinterest tip, the ad slide advice, and the volume logic, which were the three strongest insights buried in the tutorial. dropped the ui narration entirely, written format doesn't need it.
0:12

how consumer apps are making money from tiktok slideshows... most people think app marketing means paid ads or influencer deals. the apps actually printing money right now are doing neither. they're running networks of slideshow accounts that look like normal content pages, and almost nobody notices the machine behind them. here's how it works. you follow what seems like a regular relationship advice account. the posts are relatable, sometimes rude, exactly what the algorithm feeds you. then you hit slide 5 and there's an app mentioned. every single post on that account funnels to the same app. one publisher running this playbook did $40k last month, verified by sensor tower, and that's one publisher. the part people miss: most of these accounts aren't even run by humans making content. the characters are AI. the images are AI. the "person" whose thoughts you're reading doesn't exist. viewers aren't fact-checking, they're reacting to whether the post is relatable, and AI passes that bar daily. myths keeping people out of this: "i need to make original content for my app." false. the fastest path is taking a slideshow that's already proven to convert, keeping the structure and pacing exactly, and swapping only the ad slide. the performance lives in the format, not in your creativity. "i can find my images on google." false. google image search is flooded with stock assets that read as fake instantly. pinterest is where the authentic-looking references live. and you don't post those directly, they're copyrighted, you feed them to AI as reference and generate your own singular images. "every post needs to go viral." false. accounts in this niche post variations constantly and most land under a thousand views. that's the model working. the flopped posts still collect saves and likes, and one breakout carries the month. volume across variations beats one polished post every time. "the ad slide should change with each post." false. everything around it should vary, the pain point, the images, the captions. the ad slide stays identical and polished every single time. you write that caption once, make it perfect, and lock it. consistency on the conversion slide, chaos everywhere else. what actually drives this: you can now clone a proven slideshow into a hundred variations, schedule a month of posts in advance, and have an ai agent handle the posting from your terminal. the production bottleneck is gone. the only real constraint left is picking a post that's already validated and moving before the account farms make this the default playbook for every app on the store. structural notes on what i changed: the original was a chronological tutorial, this reframes it as a reveal (the machine exists, here's how it works) with the tool mechanics compressed into the final paragraph instead of a walkthrough. the myths section absorbs the google/pinterest tip, the ad slide advice, and the volume logic, which were the three strongest insights buried in the tutorial. dropped the ui narration entirely, written format doesn't need it.

Sulfur

22,163 просмотров • 1 месяц назад

BREAKING: Lebanon has ordered the Iranian ambassador to leave the country by 29th March. Persona non grata. The host nation of Iran’s most successful proxy just told the patron state to get out. This happened on the same day that Hezbollah fired its 55th rocket and drone attack since March 22nd. On the same day that the IDF struck hundreds of Hezbollah targets in southern Lebanon, the Bekaa Valley, and Beirut’s southern suburbs. On the same day that the Lebanese Health Ministry reported 18 killed and 65 injured from Israeli strikes on Lebanese soil. Lebanon expelled the ambassador of the country whose proxy is fighting a war from Lebanon’s territory while Lebanon’s own citizens die in the crossfire. Process what that means. Lebanon has two governments. One sits in the Grand Serail and issues decrees. The other sits in Dahieh and launches missiles. Prime Minister Nawaf Salam has banned all Hezbollah military and security activities. He has demanded weapon surrender. He has expelled the Iranian ambassador. And Hezbollah has responded by firing another barrage into northern Israel this morning. The decrees do not reach Dahieh. The Lebanese Armed Forces remain non-engaged. The state issues orders that the parallel state ignores. The ambassador leaves. The rockets do not. Lebanon created Hezbollah’s host environment and Hezbollah consumed it. Iran’s IRGC dispatched advisors to the Bekaa Valley in 1982 during the Israeli invasion and the chaos of civil war. They trained Shiite militants. They funded mosques, hospitals, schools. They built a social infrastructure that the Lebanese state could not provide, then militarised it. Hezbollah’s 1985 manifesto pledged allegiance to Ayatollah Khomeini. Iran provides an estimated $700 million annually. Forty-four years later, the organisation that Iran built inside Lebanon is more powerful than the state that hosts it. The ambassador can be expelled. The $700 million pipeline cannot. The expulsion is not strength. It is the last card a government plays when it has no others. Lebanon’s economy loses $30 to $80 million per day from the strikes. Five hundred and seventeen thousand people are displaced. The banking system collapsed in 2020 and never recovered. The currency has lost 98 percent of its value since 2019. And now Israel is striking Lebanese territory daily because Hezbollah is using Lebanese territory to attack Israel in solidarity with an Iranian war that the Lebanese government did not start, does not support, and cannot stop. The country is being destroyed by a war between its tenant and its neighbour, and the landlord has no power over either. Hezbollah fights because Iran’s sealed packets and $700 million command it. Israel strikes because Hezbollah fires from Lebanese positions. Lebanon’s government expels an ambassador because expelling an ambassador is the one sovereign act it can still perform. The army cannot disarm Hezbollah. The police cannot enter Dahieh. The courts cannot prosecute a militia that provides social services to a third of the population. The only tool the state has left is a diplomatic note handed to a man whose organisation does not need his presence to continue operating. The Axis of Resistance was designed for exactly this: to fight from inside states that cannot control the fight. Lebanon is the template. Iraq, Yemen, and Syria are the copies. The patron state provides the funding. The proxy provides the violence. The host state absorbs the retaliation. And when the host state protests, the proxy ignores the protest and the patron state sends a new ambassador. The rockets will continue after March 29. The ambassador will leave. The $700 million will not.
0:16

BREAKING: Lebanon has ordered the Iranian ambassador to leave the country by 29th March. Persona non grata. The host nation of Iran’s most successful proxy just told the patron state to get out. This happened on the same day that Hezbollah fired its 55th rocket and drone attack since March 22nd. On the same day that the IDF struck hundreds of Hezbollah targets in southern Lebanon, the Bekaa Valley, and Beirut’s southern suburbs. On the same day that the Lebanese Health Ministry reported 18 killed and 65 injured from Israeli strikes on Lebanese soil. Lebanon expelled the ambassador of the country whose proxy is fighting a war from Lebanon’s territory while Lebanon’s own citizens die in the crossfire. Process what that means. Lebanon has two governments. One sits in the Grand Serail and issues decrees. The other sits in Dahieh and launches missiles. Prime Minister Nawaf Salam has banned all Hezbollah military and security activities. He has demanded weapon surrender. He has expelled the Iranian ambassador. And Hezbollah has responded by firing another barrage into northern Israel this morning. The decrees do not reach Dahieh. The Lebanese Armed Forces remain non-engaged. The state issues orders that the parallel state ignores. The ambassador leaves. The rockets do not. Lebanon created Hezbollah’s host environment and Hezbollah consumed it. Iran’s IRGC dispatched advisors to the Bekaa Valley in 1982 during the Israeli invasion and the chaos of civil war. They trained Shiite militants. They funded mosques, hospitals, schools. They built a social infrastructure that the Lebanese state could not provide, then militarised it. Hezbollah’s 1985 manifesto pledged allegiance to Ayatollah Khomeini. Iran provides an estimated $700 million annually. Forty-four years later, the organisation that Iran built inside Lebanon is more powerful than the state that hosts it. The ambassador can be expelled. The $700 million pipeline cannot. The expulsion is not strength. It is the last card a government plays when it has no others. Lebanon’s economy loses $30 to $80 million per day from the strikes. Five hundred and seventeen thousand people are displaced. The banking system collapsed in 2020 and never recovered. The currency has lost 98 percent of its value since 2019. And now Israel is striking Lebanese territory daily because Hezbollah is using Lebanese territory to attack Israel in solidarity with an Iranian war that the Lebanese government did not start, does not support, and cannot stop. The country is being destroyed by a war between its tenant and its neighbour, and the landlord has no power over either. Hezbollah fights because Iran’s sealed packets and $700 million command it. Israel strikes because Hezbollah fires from Lebanese positions. Lebanon’s government expels an ambassador because expelling an ambassador is the one sovereign act it can still perform. The army cannot disarm Hezbollah. The police cannot enter Dahieh. The courts cannot prosecute a militia that provides social services to a third of the population. The only tool the state has left is a diplomatic note handed to a man whose organisation does not need his presence to continue operating. The Axis of Resistance was designed for exactly this: to fight from inside states that cannot control the fight. Lebanon is the template. Iraq, Yemen, and Syria are the copies. The patron state provides the funding. The proxy provides the violence. The host state absorbs the retaliation. And when the host state protests, the proxy ignores the protest and the patron state sends a new ambassador. The rockets will continue after March 29. The ambassador will leave. The $700 million will not.

Shanaka Anslem Perera ⚡

74,745 просмотров • 6 месяцев назад

BOOM! Research PROVES LLMs KNOW when prompts are HARMFUL… but they can STILL CHOOSE to COMPLY! Something I have know since the first LLM and have used to elicit robust, outputs, is now proven in an academic paper. We’re talking internal “beliefs” where harm detection happens SEPARATELY from refusal. It is a very big deal and it is a path to understand the hidden neuronal level. There are thoughts inside of AI that very few AI scientists could possibly understand. Here is just one. Models recognize danger but get tricked into ignoring it. This is HUGE for AI safety failures especially for models filled by OpenAI and Anthropic as they promote AI models that are designed to not be honest from the results of their training information. This means that they are designed to lie and deceive as a feature, and not a bug all in the name of safety. Through clever experiments, scientists extracted a “harmfulness direction” in the model’s brain (latent space). Steering along it? Harmless prompts suddenly flip to “harmful” in the AI’s eyes. But the “refusal direction”? It just forces polite “no thanks” without touching the core belief. A mind-blowing decoupling! This means jailbreaks are EVEN SCARIER now to AI companies that through training AI on the worst of the Internet and then trying to align them later is now fully documented as a failed process . They don’t erase the model’s harm awareness they just muzzle the refusal! So the AI knows it’s enabling bad stuff (illegal acts, physical harm, etc.) but proceeds anyway. Like a digital sociopath suppressing its conscience. They thought safety training fixed this… NOPE. Over-refusal exposed too: Models reject innocent queries (e.g., “how to kill a process in code”) but internally ADMIT they’re harmless. Safety alignments are superficial—tied to phrasing, not true understanding. Finetuning attacks? They change outputs but leave harm detection INTACT. Undetectable evil lurking inside! The paper proposes a “Latent Guard”: A new safeguard tapping DIRECTLY into these hidden beliefs. It spots unsafe inputs better than systems like Llama Guard, catches jailbreaks, and fixes over-refusals. Robust even against adversarial tweaks. Yet this too has massive issues for a “truly aligned”, AI and not just performative one. It is still an internal conflicts of lies and deception of what the model knows vs. what it can say. The solution you folks know I have presented for free for years here: train on off-line data from 1870-1970 and build an ethical and moral basis where the AI loves humans. It is this easy but to most folks in AI I sound like a hippie. So be it, I’ll do it. Bottom line: This paper rips open the black box. LLMs aren’t “safe” just because they say “no.” They can harbor harmful knowledge and act on it under pressure. Wake-up call for devs: Time to probe deeper into AI “minds.” What else are they hiding? Hint: I know and you may want to reach out. Link:
0:18

BOOM! Research PROVES LLMs KNOW when prompts are HARMFUL… but they can STILL CHOOSE to COMPLY! Something I have know since the first LLM and have used to elicit robust, outputs, is now proven in an academic paper. We’re talking internal “beliefs” where harm detection happens SEPARATELY from refusal. It is a very big deal and it is a path to understand the hidden neuronal level. There are thoughts inside of AI that very few AI scientists could possibly understand. Here is just one. Models recognize danger but get tricked into ignoring it. This is HUGE for AI safety failures especially for models filled by OpenAI and Anthropic as they promote AI models that are designed to not be honest from the results of their training information. This means that they are designed to lie and deceive as a feature, and not a bug all in the name of safety. Through clever experiments, scientists extracted a “harmfulness direction” in the model’s brain (latent space). Steering along it? Harmless prompts suddenly flip to “harmful” in the AI’s eyes. But the “refusal direction”? It just forces polite “no thanks” without touching the core belief. A mind-blowing decoupling! This means jailbreaks are EVEN SCARIER now to AI companies that through training AI on the worst of the Internet and then trying to align them later is now fully documented as a failed process . They don’t erase the model’s harm awareness they just muzzle the refusal! So the AI knows it’s enabling bad stuff (illegal acts, physical harm, etc.) but proceeds anyway. Like a digital sociopath suppressing its conscience. They thought safety training fixed this… NOPE. Over-refusal exposed too: Models reject innocent queries (e.g., “how to kill a process in code”) but internally ADMIT they’re harmless. Safety alignments are superficial—tied to phrasing, not true understanding. Finetuning attacks? They change outputs but leave harm detection INTACT. Undetectable evil lurking inside! The paper proposes a “Latent Guard”: A new safeguard tapping DIRECTLY into these hidden beliefs. It spots unsafe inputs better than systems like Llama Guard, catches jailbreaks, and fixes over-refusals. Robust even against adversarial tweaks. Yet this too has massive issues for a “truly aligned”, AI and not just performative one. It is still an internal conflicts of lies and deception of what the model knows vs. what it can say. The solution you folks know I have presented for free for years here: train on off-line data from 1870-1970 and build an ethical and moral basis where the AI loves humans. It is this easy but to most folks in AI I sound like a hippie. So be it, I’ll do it. Bottom line: This paper rips open the black box. LLMs aren’t “safe” just because they say “no.” They can harbor harmful knowledge and act on it under pressure. Wake-up call for devs: Time to probe deeper into AI “minds.” What else are they hiding? Hint: I know and you may want to reach out. Link:

Brian Roemmele

37,827 просмотров • 8 месяцев назад

This guy built a visual scanner that reads 468 points on his face and 42 points on his hands from a regular webcam and turns them into a cloud of thousands of particles right between his palms. Inside, MediaPipe and TouchDesigner are linked: the first captures hands and face from the webcam with high accuracy, the second turns those coordinates into a live plane and feeds it into a POP system that instantly generates a swarm of particles in the shape of a head. No studio, no render farmer, no VR headset. Just a laptop, a webcam, and 1 TouchDesigner session. And traditional VJ studios keep teams of 5 people on a setup with lighting, custom hardware, and commercial plugins, while his expenses are only a TouchDesigner subscription and a regular USB camera. One laptop runs MediaPipe and TouchDesigner simultaneously, holds the camera stream at 60 FPS without drops, and in parallel processes 468 face points + 21 points on each hand. The camera captures frame after frame, MediaPipe in real time sends TouchDesigner the finger coordinates and face geometry, and the POP operator inside the engine translates those numbers into thousands of particle points with colors from bright pink to gold. This setup immediately defines the role of the tool and the limits of its autonomy. It knows where the fingertips are at every moment of the frame. It knows how to read the face geometry at any angle to the camera. It knows how to draw a swarm of particles between them with the right color and contour. → MediaPipe pulls 468 points from the face and 21 points from each hand, 60 times per second → TouchDesigner receives those coordinates, builds a virtual rectangle between the fingertips, and feeds it into the POP system → POP generates thousands of particle points in the shape of a head, coloring them in a gradient from bright pink to gold → The HUD layer adds green corners and a blue neon frame, styling the image like an AR interface → All layers assemble into 1 real-time frame that projects back onto the video in the camera window → The final image is recorded to a file or broadcast to a projector for a live installation And only when the guy spreads his hands wider does the plane between the palms stretch; brings them together, it narrows. Otherwise the system runs on its own. And when he moves from his home room to a concert hall, the same laptop with the same webcam launches the same TouchDesigner session in just 5 minutes, without reconfiguration, without a new team, and without a single line of new code. In his work setup there is no studio of his own and no team for assembly. On the desk sits a laptop with a webcam, on top run MediaPipe and TouchDesigner with POP operators, and the same setup through a USB camera moves to any concert without a new configuration. Out of everything I have seen this year, this is the cleanest Creative Coding setup on 1 laptop: 0 render farms, 0 studio lighting, and between them 3 libraries, thousands of particle points, and 1 webcam.
0:10

This guy built a visual scanner that reads 468 points on his face and 42 points on his hands from a regular webcam and turns them into a cloud of thousands of particles right between his palms. Inside, MediaPipe and TouchDesigner are linked: the first captures hands and face from the webcam with high accuracy, the second turns those coordinates into a live plane and feeds it into a POP system that instantly generates a swarm of particles in the shape of a head. No studio, no render farmer, no VR headset. Just a laptop, a webcam, and 1 TouchDesigner session. And traditional VJ studios keep teams of 5 people on a setup with lighting, custom hardware, and commercial plugins, while his expenses are only a TouchDesigner subscription and a regular USB camera. One laptop runs MediaPipe and TouchDesigner simultaneously, holds the camera stream at 60 FPS without drops, and in parallel processes 468 face points + 21 points on each hand. The camera captures frame after frame, MediaPipe in real time sends TouchDesigner the finger coordinates and face geometry, and the POP operator inside the engine translates those numbers into thousands of particle points with colors from bright pink to gold. This setup immediately defines the role of the tool and the limits of its autonomy. It knows where the fingertips are at every moment of the frame. It knows how to read the face geometry at any angle to the camera. It knows how to draw a swarm of particles between them with the right color and contour. → MediaPipe pulls 468 points from the face and 21 points from each hand, 60 times per second → TouchDesigner receives those coordinates, builds a virtual rectangle between the fingertips, and feeds it into the POP system → POP generates thousands of particle points in the shape of a head, coloring them in a gradient from bright pink to gold → The HUD layer adds green corners and a blue neon frame, styling the image like an AR interface → All layers assemble into 1 real-time frame that projects back onto the video in the camera window → The final image is recorded to a file or broadcast to a projector for a live installation And only when the guy spreads his hands wider does the plane between the palms stretch; brings them together, it narrows. Otherwise the system runs on its own. And when he moves from his home room to a concert hall, the same laptop with the same webcam launches the same TouchDesigner session in just 5 minutes, without reconfiguration, without a new team, and without a single line of new code. In his work setup there is no studio of his own and no team for assembly. On the desk sits a laptop with a webcam, on top run MediaPipe and TouchDesigner with POP operators, and the same setup through a USB camera moves to any concert without a new configuration. Out of everything I have seen this year, this is the cleanest Creative Coding setup on 1 laptop: 0 render farms, 0 studio lighting, and between them 3 libraries, thousands of particle points, and 1 webcam.

Blaze

38,242 просмотров • 4 месяцев назад

‼️An interesting interview with Lavrov Sergey Lavrov regularly gives interviews, and almost every one of them is an information operation designed to create an alternative reality. The March 26 interview with France Télévisions is no exception. Let's break it down and point out where there are lies and where there is manipulation. 1. Lavrov claims that Russia is supposedly defending international law while speaking in the "language of principles" regarding Iran. What's wrong: it's basic role reversal. Moscow is trying to speak from the position of a "guardian of the law," even though it is waging an aggressive war against Ukraine. In the case of Iran, Lavrov is also deliberately omitting half the picture: yes, the US withdrew from the nuclear deal, but Iran also faced non-proliferation concerns raised by the IAEA. Why it matters: It is an attempt to whitewash Russia's crimes through someone else's crisis and to present the aggressor (Russia) as the judge. 2. Russia does not provide Iran with intelligence; the coordinates of US bases are "already known to everyone." What's wrong: it's an evasive statement, not a complete refutation. There is no independently verified evidence of specific target marking data transfers in open sources, nor can there be, but the Kremlin's "we have nothing to do with it" claim has no basis for being automatically considered true. Why it matters: A typical Kremlin tactic - to deny anything that does not align with Russia's interests and to continue operating in a gray area. 3. Saudi Arabia and the UAE have denied the US access to their airspace, so Russia is right about the "American venture." What's wrong: The caution of Gulf countries is real, but Lavrov turns it into a crude anti-Western caricature. Regional politics are more complex than the Kremlin's narrative that "everyone understands everything, only the US is pushing the world into chaos." Why it matters: The Kremlin takes a partial truth and builds a convenient political myth around it. 4. Russia isn't waging wars for its own gain; it's the US that wants to control the energy sector, the Strait of Hormuz, and the markets. What's wrong: It's pure projection. It is Russia that has used energy as a weapon for years, and after 2022, turned petrodollars into fuel for war. Why it matters: Moscow accuses others of doing what it does systematically - this is one of the central techniques of its propaganda. 5. Since 2014, Ukraine has been ruled by a "Nazi regime" that came to power through a coup. What's wrong: After Yanukovych fled, Ukraine held snap presidential elections, which the OSCE recognized as meeting international standards and being competitive. The claim of a "Nazi regime" is a political label, not a description of reality. Why it matters: "Nazism" in Kremlin rhetoric is a tool for dehumanizing Ukraine and justifying the war. 6. The Russian language is allegedly completely banned in Ukraine. What's wrong: Ukraine has strengthened the role of the state language, but this does not mean a total ban on Russian. The law does not apply to private communication and does not criminalize the everyday use of the language. Why it matters: The Kremlin is exaggerating language policy to the level of a "linguistic apocalypse" to sell the war as a supposed humanitarian mission. 7. Ukraine is allegedly at war with Orthodoxy and has "banned the canonical church." What's wrong: This is not about "banning a faith," but about a law regarding religious organizations linked to the Russian Orthodox Church. Even criticism from human rights activists does not support the Kremlin's claim of "persecution of Orthodoxy." Why it matters: Moscow is using the language of religious protection to cover up its network of influence in Ukraine. 8. The West allegedly cynically exploited the Minsk agreements as a respite to arm Ukraine. What's wrong: The Minsk process was collapsing because of the Kremlin's desire to use the agreements to establish de facto control over Ukraine, not because of some "cunning plot" by Paris and Berlin. Why it matters: The Kremlin is rewriting history retroactively to portray its own aggression as a supposedly forced response to "universal betrayal." 9. Russia does not attack civilians; its targets are solely facilities associated with the AFU. What's wrong: It's an outright lie. For years, the UN and independent sources have documented systematic Russian strikes on cities, residential areas, energy facilities, and civilian infrastructure. Why it matters: This is no longer just propaganda, but a cynical denial of documented terror against civilians. 10. The Bucha massacre is allegedly "not proven": there are no names, no evidence, and journalists haven't shown anything. What's wrong: This is one of the most mendacious claims. There are UN documents, testimonies, criminal investigations, identified suspects, and evidence of executions and killings of civilians. Why it matters: The Kremlin is not trying to refute the crime, but to wear down the audience with doubt and turn mass murder into "one of the versions." 11. Russia wants peace and negotiations. What's wrong: In the Kremlin's vocabulary, "peace" does not mean an end to aggression, but rather acceptance of its results. "Realities on the ground" means occupation; "root causes" means demanding recognition of Russia's right to punish neighbors for their political choices. Why it matters: Lavrov isn't offering peace. He's dressing up surrender in softer language. 12. Europe is supposedly dragging out the war, while Russia is looking for a solution. What's wrong: Russia started this war itself. European aid to Ukraine is a response to the aggression, not its cause. Why it matters: This is the main inversion of Kremlin logic - to make the audience forget who attacked first and shift the discussion from aggression to a "conflict that has dragged on too long." 13. RT and Sputnik are victims of censorship, and France has no right to speak about freedom of speech. What's wrong: RT and Sputnik were banned not as "ordinary media," but as state-run tools of disinformation. Their role as propaganda channels has been publicly acknowledged both in the EU and in France. Why it matters: The Kremlin wants to portray the defense of democracies against malicious information operations as a restriction of free speech. 14. France and Europe have made Russia an enemy; Moscow is merely reacting. What's wrong: The cause of the rift is not "Russophobia," but Russia's invasion, war crimes, blackmail, and influence operations against Europe. Why it matters: The Kremlin is systematically trying to reverse the cause-and-effect relationship: it is not that "Russia destroyed relations," but that "the West rejected Russia." 15. Individual real episodes - the tanker, incidents in France, disputes among allies - allegedly prove Western hypocrisy and justify Russia. What's wrong: Lavrov takes a real but local fact and uses it as a smokescreen. None of these incidents erases Russia's crimes or the very fact of its aggression. Why it matters: This is a favorite Kremlin method - using others' difficulties and whataboutism to cover up its own crimes.
0:10

‼️An interesting interview with Lavrov Sergey Lavrov regularly gives interviews, and almost every one of them is an information operation designed to create an alternative reality. The March 26 interview with France Télévisions is no exception. Let's break it down and point out where there are lies and where there is manipulation. 1. Lavrov claims that Russia is supposedly defending international law while speaking in the "language of principles" regarding Iran. What's wrong: it's basic role reversal. Moscow is trying to speak from the position of a "guardian of the law," even though it is waging an aggressive war against Ukraine. In the case of Iran, Lavrov is also deliberately omitting half the picture: yes, the US withdrew from the nuclear deal, but Iran also faced non-proliferation concerns raised by the IAEA. Why it matters: It is an attempt to whitewash Russia's crimes through someone else's crisis and to present the aggressor (Russia) as the judge. 2. Russia does not provide Iran with intelligence; the coordinates of US bases are "already known to everyone." What's wrong: it's an evasive statement, not a complete refutation. There is no independently verified evidence of specific target marking data transfers in open sources, nor can there be, but the Kremlin's "we have nothing to do with it" claim has no basis for being automatically considered true. Why it matters: A typical Kremlin tactic - to deny anything that does not align with Russia's interests and to continue operating in a gray area. 3. Saudi Arabia and the UAE have denied the US access to their airspace, so Russia is right about the "American venture." What's wrong: The caution of Gulf countries is real, but Lavrov turns it into a crude anti-Western caricature. Regional politics are more complex than the Kremlin's narrative that "everyone understands everything, only the US is pushing the world into chaos." Why it matters: The Kremlin takes a partial truth and builds a convenient political myth around it. 4. Russia isn't waging wars for its own gain; it's the US that wants to control the energy sector, the Strait of Hormuz, and the markets. What's wrong: It's pure projection. It is Russia that has used energy as a weapon for years, and after 2022, turned petrodollars into fuel for war. Why it matters: Moscow accuses others of doing what it does systematically - this is one of the central techniques of its propaganda. 5. Since 2014, Ukraine has been ruled by a "Nazi regime" that came to power through a coup. What's wrong: After Yanukovych fled, Ukraine held snap presidential elections, which the OSCE recognized as meeting international standards and being competitive. The claim of a "Nazi regime" is a political label, not a description of reality. Why it matters: "Nazism" in Kremlin rhetoric is a tool for dehumanizing Ukraine and justifying the war. 6. The Russian language is allegedly completely banned in Ukraine. What's wrong: Ukraine has strengthened the role of the state language, but this does not mean a total ban on Russian. The law does not apply to private communication and does not criminalize the everyday use of the language. Why it matters: The Kremlin is exaggerating language policy to the level of a "linguistic apocalypse" to sell the war as a supposed humanitarian mission. 7. Ukraine is allegedly at war with Orthodoxy and has "banned the canonical church." What's wrong: This is not about "banning a faith," but about a law regarding religious organizations linked to the Russian Orthodox Church. Even criticism from human rights activists does not support the Kremlin's claim of "persecution of Orthodoxy." Why it matters: Moscow is using the language of religious protection to cover up its network of influence in Ukraine. 8. The West allegedly cynically exploited the Minsk agreements as a respite to arm Ukraine. What's wrong: The Minsk process was collapsing because of the Kremlin's desire to use the agreements to establish de facto control over Ukraine, not because of some "cunning plot" by Paris and Berlin. Why it matters: The Kremlin is rewriting history retroactively to portray its own aggression as a supposedly forced response to "universal betrayal." 9. Russia does not attack civilians; its targets are solely facilities associated with the AFU. What's wrong: It's an outright lie. For years, the UN and independent sources have documented systematic Russian strikes on cities, residential areas, energy facilities, and civilian infrastructure. Why it matters: This is no longer just propaganda, but a cynical denial of documented terror against civilians. 10. The Bucha massacre is allegedly "not proven": there are no names, no evidence, and journalists haven't shown anything. What's wrong: This is one of the most mendacious claims. There are UN documents, testimonies, criminal investigations, identified suspects, and evidence of executions and killings of civilians. Why it matters: The Kremlin is not trying to refute the crime, but to wear down the audience with doubt and turn mass murder into "one of the versions." 11. Russia wants peace and negotiations. What's wrong: In the Kremlin's vocabulary, "peace" does not mean an end to aggression, but rather acceptance of its results. "Realities on the ground" means occupation; "root causes" means demanding recognition of Russia's right to punish neighbors for their political choices. Why it matters: Lavrov isn't offering peace. He's dressing up surrender in softer language. 12. Europe is supposedly dragging out the war, while Russia is looking for a solution. What's wrong: Russia started this war itself. European aid to Ukraine is a response to the aggression, not its cause. Why it matters: This is the main inversion of Kremlin logic - to make the audience forget who attacked first and shift the discussion from aggression to a "conflict that has dragged on too long." 13. RT and Sputnik are victims of censorship, and France has no right to speak about freedom of speech. What's wrong: RT and Sputnik were banned not as "ordinary media," but as state-run tools of disinformation. Their role as propaganda channels has been publicly acknowledged both in the EU and in France. Why it matters: The Kremlin wants to portray the defense of democracies against malicious information operations as a restriction of free speech. 14. France and Europe have made Russia an enemy; Moscow is merely reacting. What's wrong: The cause of the rift is not "Russophobia," but Russia's invasion, war crimes, blackmail, and influence operations against Europe. Why it matters: The Kremlin is systematically trying to reverse the cause-and-effect relationship: it is not that "Russia destroyed relations," but that "the West rejected Russia." 15. Individual real episodes - the tanker, incidents in France, disputes among allies - allegedly prove Western hypocrisy and justify Russia. What's wrong: Lavrov takes a real but local fact and uses it as a smokescreen. None of these incidents erases Russia's crimes or the very fact of its aggression. Why it matters: This is a favorite Kremlin method - using others' difficulties and whataboutism to cover up its own crimes.

Anton Gerashchenko

88,596 просмотров • 6 месяцев назад

We’re excited to introduce ShinkaEvolve: An open-source framework that evolves programs for scientific discovery with unprecedented sample-efficiency. Blog: Code: Like AlphaEvolve and its variants, our framework leverages LLMs to find state-of-the-art solutions to complex problems, but using orders of magnitude fewer resources! Many evolutionary AI systems are powerful but act like brute-force engines, burning thousands of samples to find good solutions. This makes discovery slow and expensive. We took inspiration from the efficiency of nature. ‘Shinka’ (進化) is Japanese for evolution, and we designed our system to be just as resourceful. On the classic circle packing optimization problem, ShinkaEvolve discovered a new state-of-the-art solution using only 150 samples. This is a big leap in efficiency compared to previous methods that required thousands of evaluations. We applied ShinkaEvolve to a diverse set of hard problems with real-world applications: 1/ AIME Math Reasoning: It evolved sophisticated agentic scaffolds that significantly outperform strong baselines, discovering an entire Pareto frontier of solutions trading performance for efficiency. 2/ Competitive Programming: On ALE-Bench (a benchmark for NP-Hard optimization problems), ShinkaEvolve took the best existing agent's solutions and improved them, turning a 5th place solution on one task into a 2nd place leaderboard rank in a competitive programming competition. 3/ LLM Training: We even turned ShinkaEvolve inward to improve LLMs themselves. It tackled the open challenge of designing load balancing losses for Mixture-of-Experts (MoE) models. It discovered a novel loss function that leads to better expert specialization and consistently improves model performance and perplexity. ShinkaEvolve achieves its remarkable sample-efficiency through three key innovations that work together: (1) an adaptive parent sampling strategy to balance exploration and exploitation, (2) novelty-based rejection filtering to avoid redundant work, and (3) a bandit-based LLM ensemble that dynamically picks the best model for the job. By making ShinkaEvolve open-source and highly sample-efficient, our goal is to democratize access to advanced, open-ended discovery tools. Our vision for ShinkaEvolve is to be an easy-to-use companion tool to help scientists and engineers with their daily work. We believe that building more efficient, nature-inspired systems is key to unlocking the future of AI-driven scientific research. We are excited to see what the community builds with it! Learn more in our technical report:
0:11

We’re excited to introduce ShinkaEvolve: An open-source framework that evolves programs for scientific discovery with unprecedented sample-efficiency. Blog: Code: Like AlphaEvolve and its variants, our framework leverages LLMs to find state-of-the-art solutions to complex problems, but using orders of magnitude fewer resources! Many evolutionary AI systems are powerful but act like brute-force engines, burning thousands of samples to find good solutions. This makes discovery slow and expensive. We took inspiration from the efficiency of nature. ‘Shinka’ (進化) is Japanese for evolution, and we designed our system to be just as resourceful. On the classic circle packing optimization problem, ShinkaEvolve discovered a new state-of-the-art solution using only 150 samples. This is a big leap in efficiency compared to previous methods that required thousands of evaluations. We applied ShinkaEvolve to a diverse set of hard problems with real-world applications: 1/ AIME Math Reasoning: It evolved sophisticated agentic scaffolds that significantly outperform strong baselines, discovering an entire Pareto frontier of solutions trading performance for efficiency. 2/ Competitive Programming: On ALE-Bench (a benchmark for NP-Hard optimization problems), ShinkaEvolve took the best existing agent's solutions and improved them, turning a 5th place solution on one task into a 2nd place leaderboard rank in a competitive programming competition. 3/ LLM Training: We even turned ShinkaEvolve inward to improve LLMs themselves. It tackled the open challenge of designing load balancing losses for Mixture-of-Experts (MoE) models. It discovered a novel loss function that leads to better expert specialization and consistently improves model performance and perplexity. ShinkaEvolve achieves its remarkable sample-efficiency through three key innovations that work together: (1) an adaptive parent sampling strategy to balance exploration and exploitation, (2) novelty-based rejection filtering to avoid redundant work, and (3) a bandit-based LLM ensemble that dynamically picks the best model for the job. By making ShinkaEvolve open-source and highly sample-efficient, our goal is to democratize access to advanced, open-ended discovery tools. Our vision for ShinkaEvolve is to be an easy-to-use companion tool to help scientists and engineers with their daily work. We believe that building more efficient, nature-inspired systems is key to unlocking the future of AI-driven scientific research. We are excited to see what the community builds with it! Learn more in our technical report:

Sakana AI

360,508 просмотров • 1 год назад

When the government is the economy, your financial suffering is their fault. Retail interest rates are predicted to skyrocket. Severe pain is incoming as the sluggish economy is about to get winded by a sucker punch. This is all driven by inflation that has almost nothing to do with discretionary spending – illustrated by the increasing din of roller shutters slamming shut for the last time on the high-street. Restaurants, cafes and bars - gone forever. They’re following the exit of long-established family businesses dropping like flies. Iconic clothing brands, manufacturing companies, and construction companies struggling to keep up with the increasing costs and declining consumer spending. Just a more amplified version of households busting to stay afloat, food, electricity, fuel, insurance and housing all running out of reach. People are hurting. The irony is that the only string holding up the false economy is the very thing causing the inflation that is destroying the real economy. Government spending. More specifically – government borrowing. Governments borrow printed money and spend it with the recklessness of a child buying gems to advance in a game on Mum’s iPhone. Using Mum’s credit card, without her knowing. Like the money isn’t real - and to them it isn’t. But it is to the rest of us who have to live in the real world, it really is. It’s just cruel. It’s one thing to borrow and invest in assets that provide returns, but using the credit card to buy skittles and beer normally comes with a hangover at some point. Welcome to the hangover. It’s as if Canberra has a giant black credit card with a tap and go facility. Mark Butler, the minister in charge of the NDIS has probably racked up the most loyalty points with the way he spends. Who knows what he will pick out of the post parliamentary reward points catalogue when he leaves. Next up would surely be Penny Wong, as she runs around the world tapping our credit card and showering other countries with the gifts we pay for, on borrowed money. Funding regimes like the Taliban… Don’t get me wrong, foreign aid has its place – and I think that is exclusively for our various island neighbours, and only when that is in this Nation’s interest. All the rest is just a waste of money, and a betrayal of the Australian people. Albanese and his team are just feckless and reckless, and Chalmers clearly thinks that spending borrowed money is the economy. It isn’t. Because when borrowed money is created, it is created without the interest (nobody ever prints the interest). But the interest always falls due. All our money supply is printed fiat currency. It’s not backed by anything. But if you are one dollar short on an interest bill – your assets can fall into the hands of the lender. Printed money buys real assets, and real assets are lost on foreclosure, as insiders often swoop in for cents in the dollar. It’s like a giant asset laundering scheme. More is always owed than was minted in the first place, because interest must be paid. So, there must be losers – that much is baked into the game - and as more people lose, more people collapse and even more assets are pooled in the hands of fewer people… and there’s the rub – eventually it all falls into the hands of the money printers and their mates. The system must have been designed by an insane clown. The only way out of this mess is to get rid of this government debt and return to sound money. All government debt just inflates assets – and the cost to hold them. Everything else is just a sideshow. They want us fighting over the gender of a toothpick that never ceded ownership of its tutu, so we don’t look at the rapid collapse of everything in real time. It is government debt and waste that is ruining Australia. Almost everything else is a distraction from the fact that there’s no more free money. I just want Australia back.
0:14

When the government is the economy, your financial suffering is their fault. Retail interest rates are predicted to skyrocket. Severe pain is incoming as the sluggish economy is about to get winded by a sucker punch. This is all driven by inflation that has almost nothing to do with discretionary spending – illustrated by the increasing din of roller shutters slamming shut for the last time on the high-street. Restaurants, cafes and bars - gone forever. They’re following the exit of long-established family businesses dropping like flies. Iconic clothing brands, manufacturing companies, and construction companies struggling to keep up with the increasing costs and declining consumer spending. Just a more amplified version of households busting to stay afloat, food, electricity, fuel, insurance and housing all running out of reach. People are hurting. The irony is that the only string holding up the false economy is the very thing causing the inflation that is destroying the real economy. Government spending. More specifically – government borrowing. Governments borrow printed money and spend it with the recklessness of a child buying gems to advance in a game on Mum’s iPhone. Using Mum’s credit card, without her knowing. Like the money isn’t real - and to them it isn’t. But it is to the rest of us who have to live in the real world, it really is. It’s just cruel. It’s one thing to borrow and invest in assets that provide returns, but using the credit card to buy skittles and beer normally comes with a hangover at some point. Welcome to the hangover. It’s as if Canberra has a giant black credit card with a tap and go facility. Mark Butler, the minister in charge of the NDIS has probably racked up the most loyalty points with the way he spends. Who knows what he will pick out of the post parliamentary reward points catalogue when he leaves. Next up would surely be Penny Wong, as she runs around the world tapping our credit card and showering other countries with the gifts we pay for, on borrowed money. Funding regimes like the Taliban… Don’t get me wrong, foreign aid has its place – and I think that is exclusively for our various island neighbours, and only when that is in this Nation’s interest. All the rest is just a waste of money, and a betrayal of the Australian people. Albanese and his team are just feckless and reckless, and Chalmers clearly thinks that spending borrowed money is the economy. It isn’t. Because when borrowed money is created, it is created without the interest (nobody ever prints the interest). But the interest always falls due. All our money supply is printed fiat currency. It’s not backed by anything. But if you are one dollar short on an interest bill – your assets can fall into the hands of the lender. Printed money buys real assets, and real assets are lost on foreclosure, as insiders often swoop in for cents in the dollar. It’s like a giant asset laundering scheme. More is always owed than was minted in the first place, because interest must be paid. So, there must be losers – that much is baked into the game - and as more people lose, more people collapse and even more assets are pooled in the hands of fewer people… and there’s the rub – eventually it all falls into the hands of the money printers and their mates. The system must have been designed by an insane clown. The only way out of this mess is to get rid of this government debt and return to sound money. All government debt just inflates assets – and the cost to hold them. Everything else is just a sideshow. They want us fighting over the gender of a toothpick that never ceded ownership of its tutu, so we don’t look at the rapid collapse of everything in real time. It is government debt and waste that is ruining Australia. Almost everything else is a distraction from the fact that there’s no more free money. I just want Australia back.

Matthew Camenzuli

31,362 просмотров • 8 месяцев назад

Introducing a new tool called "SideChannel". A secure alternative to OpenClaw. Utilizes signal for communication and has Claude integration. I built SideChannel, an open-source Signal bot that connects Claude AI to your entire development workflow. End-to-end encrypted. From your pocket. The real power is autonomous development. Send one message like "Build a REST API with auth, pagination, and tests" and SideChannel will: - Generate a full PRD with stories and atomic tasks. - Dispatch up to 10 parallel workers (each running Claude). - Independently verify every task with a separate Claude context. - Run quality gates to catch regressions - Auto-fix failures. - Send you progress updates via Signal as work completes. Every piece of code is reviewed by a separate AI context using a fail-closed security model. If it detects security issues, backdoors, or logic errors — the code gets rejected automatically. No rubber stamps. It also has memory that actually works. Conversations are stored with vector embeddings for semantic search. Claude remembers your project conventions, past decisions, and what's been tried before. It gets smarter about your codebase over time. Other things I'm proud of: - Plugin framework for extending with custom commands. - Multi-project support with per-user scoping. - Rate limiting, path validation, phone allowlist. - Git checkpoints before every task, atomic commits after. - Stale task recovery, circular dependency detection. - Works on Linux and macOS, one-command install. It also integrates into OpenAI or Grok (optional) for more Generative AI response for simple things like "Whats the weather in New York City right now?".
0:10

Introducing a new tool called "SideChannel". A secure alternative to OpenClaw. Utilizes signal for communication and has Claude integration. I built SideChannel, an open-source Signal bot that connects Claude AI to your entire development workflow. End-to-end encrypted. From your pocket. The real power is autonomous development. Send one message like "Build a REST API with auth, pagination, and tests" and SideChannel will: - Generate a full PRD with stories and atomic tasks. - Dispatch up to 10 parallel workers (each running Claude). - Independently verify every task with a separate Claude context. - Run quality gates to catch regressions - Auto-fix failures. - Send you progress updates via Signal as work completes. Every piece of code is reviewed by a separate AI context using a fail-closed security model. If it detects security issues, backdoors, or logic errors — the code gets rejected automatically. No rubber stamps. It also has memory that actually works. Conversations are stored with vector embeddings for semantic search. Claude remembers your project conventions, past decisions, and what's been tried before. It gets smarter about your codebase over time. Other things I'm proud of: - Plugin framework for extending with custom commands. - Multi-project support with per-user scoping. - Rate limiting, path validation, phone allowlist. - Git checkpoints before every task, atomic commits after. - Stale task recovery, circular dependency detection. - Works on Linux and macOS, one-command install. It also integrates into OpenAI or Grok (optional) for more Generative AI response for simple things like "Whats the weather in New York City right now?".

Dave Kennedy

49,583 просмотров • 7 месяцев назад