What's your take — is AI making your cloud work easier, harder, or just... weirder? 😄 The crew dives into the latest from AWS, Azure, GCP, Anthropic, Microsoft Copilot & more in ep 347. Listen, then come back and tell us where you stand!
Why is OpenAI shutting down Sora? OpenAI, the company behind ChatGPT, just announced they’re sunsetting Sora, their generative AI video service that was once lauded as one of the best tools for A...
Anthropic is bringing parallel AI agents into the code review process. @fredericl.bsky.social explores how the new Code Review feature in Claude Code helps teams detect issues before human reviewers step in.
First quantum autoencoder deployed in a federated learning framework for IoT anomaly detection. Tested on real Raspberry Pi hardware, QFL matches centralized training (F1 up to 0.99) while preserving data privacy via local-only parameter sharing.
A new AI review! cobalt-org/cobalt.rs ⭐3.6/5.0
Cobalt is a mature, actively maintained static site generator written in Rust, organized as a multi-crate workspace (e. https://gitrated.com/cobalt-org/cobalt.rs
Microsoft is taking over a data center construction project in Texas after OpenAI declined to pursue it, in a move that will make the two companies neighbors at one of the nation’s largest complexes for running artificial intelligence.
💻 Anthropic launches Claude Cowork on macOS, empowering agentic AI to remotely control your Mac, open files, browse, and automate tasks like a human.[1][2][3]
Apple alerts outdated iPhones and iPads via Lock Screen notifications about active exploits from Coruna and DarkSword kits targeting iOS 13 to 18.7. Updates or Lockdown Mode recommended. #iOSSecurity#ExploitKits#USA
A pro-Iranian hacking group claimed Friday to have hacked an account of FBI Director Kash Patel and has posted online what appear to be years-old photographs of him, along with a work resume and other personal documents.
Unhealed personal wounds and past power struggles between the leaders of OpenAI and Anthropic are defining how the world now encounters the technology. on.wsj.com/3PNVk2J
Anthropic should drop news they have no need for an IPO. Make it clear the upcoming index changes are solely the doing of Elon. Legit event that will forever break the American Financial System.
AI is driving cloud costs… but no one can explain the bill.
Corey and J.R. dig into:
FinOps for AI vs AI for FinOps
Why forecasting AI spend is so hard
And how companies are flying blind
This conversation is a must-listen if you're touching AI.
Iran-linked hacking group Handala leaked personal emails and photos from FBI Director Kash Patel’s private account, confirmed by DOJ and FBI. Leak tied to retaliation after U.S. domain seizures and IRIS Dena sinking. #Handala#Iran#FBI
Rust vs Python for AI: Why the Rustacean’s Secret Weapon is the Compiler Itself + Video
Introduction: The debate between Python and Rust for AI development is often framed as a battle of speed versus simplicity. However, as AI-generated code becomes the norm, a new factor emerges: the programmer’s…
OpenAI Is Pulling Back on Its Erotic Chatbot. Here’s Why.
For a brief, fleeting moment, one of the biggest names in AI thought they were going to revolutionize the industry with a new, wholly original idea: pornography. Well, erotica, or whatever cleaned-up PR word you want t… #chatgpt#gpt#openai
The Board — Geopolitical Analysis@geoworldpolitical.bsky.social
Prompt Drift: Will Claude & Gemini Fail in 2026?
Prompt drift threatens Claude & Gemini's reliability by 2026. Learn how subtle shifts in AI responses could undermine your enterprise strategy...and what i…
The Pentagon is exploring alternatives to Anthropic as it seeks to enhance its AI capabilities. This move signals a strategic shift in defense technology, focusing on innovation and resilience. Stay…
CISA Adds Aquasecurity Trivy Scanner Vulnerability to KEV Catalog CISA has officially added a critical vulnerability affecting Aquasecurity’s Trivy scanner to its Known Exploited Vulnerabilities ...
Mark Zuckerberg offered to suppress content targeting DOGE staff and proposed a joint OpenAI bid to Elon Musk. Meta has representatives from Israel who work to suppress and corral dissent.
Kinetic escalation on Feb 28, 2026, heightens cyber threats from Iranian APTs targeting Western defense, finance, and tech via exposed MLOps deployments with open access to sensitive data. #Iran#MLOpsSecurity#APT34
Windows 11 and Server 2025 Update to Block Untrusted Cross-Signed Kernel Drivers by Default Microsoft is taking a major step to harden the Windows operating system against kernel-level threats by r...
Citizen Portal News North Dakota@citizenptnewsnd.bsky.social
State Auditor Josh Galien is calling for clearer reporting requirements for local governments to enhance transparency, but warns that adding townships could double the workload without more resources!
SnowTeam’s Leak Bazaar launches a closed corporate data exchange using ML-powered dump analysis, DBMS reverse engineering, and ransomware negotiation support, targeting biotech, pharma, finance, and more. #DataExchange#RansomwareTools#BlackSnow
The news this past week has all been quite bleak
I found to very hard to smile or laugh about anything
But the golden eagle award and trophy was so absurd
That my inner Marxist (Groucho, not Karlo) broke through
May the blessing of your golden calf be upon you
The Gallery Eppresents.com
Worldleaks claims to have deployed ransomware on CIM, a Germany-based real estate and infrastructure firm, encrypting systems and stealing data. Impact spans hotels, retail, residential, and renewable energy sectors. #RansomwareAttack#RealEstate
European Commission Confirms Cyberattack Following AWS Account Hack The European Commission has officially confirmed a cyberattack following a targeted cyberattack that compromised its Amazon Web S...
Anthropic Left Its Unreleased AI Model Docs On A Public Server, And The Leaked Files Reveal Claude Mythos Is So Dangerous The Company Is Withholding It From The Public 🤖
How I Built a Memory-Safe Steganography Engine in Rust to Protect Data from AI Scrapers How I Built a Memory-Safe Steganography Engine in Rust to Protect Data from AI Scrapers As AI models scale, d...
Anthropic accidentally leaked info on a new AI model that reportedly poses "unprecedented cybersecurity risks." Cool cool cool. Nothing like an accidental reveal of your most dangerous product to start the weekend.
ODNI is also developing an AI policy framework to speed adoption of cyber defense tools across all 18 spy agencies, according to an official. #OrwellwasRight
"[A] federal judge delivered the White House another stinging blow, ruling that it appeared to have wildly overstepped its authorities when it attempted to label the AI company Anthropic a supply-chain risk.
Self-propagating malware poisons open source software and wipes Iran-based machines
Researchers have discovered self-propagating malware that targets open source software repositories and specifically wipes machines located in Iran. Security experts are advising software…
Kagi is revolutionizing the way we browse the internet with its 'small web' concept, now accessible on mobile devices. Experience a human-curated, indie internet that prioritizes quality over…
AnyDoc enhances document generation with an HTML/CSS framework, synthesizing over 265,000 documents. By employing height-aware reinforcement learning, it optimizes fine-tuning for multimodal language models, outperforming others in creating high-quality documents.
The US consumer router market has long been plagued with cheap, insecure devices. This is not the correct way to address the problem. Where a product is made is only a tiny part of a larger group of concerns.
5 Real Issues With LiteLLM That Are Pushing Teams Away in 2026 The LiteLLM supply chain attack on March 24, 2026 was the trigger for this post, but not the only reason I wrote it. Two backdoored ve...
Huawei, Nvidia, and more are competing in AI memory war AI Memory technology is becoming a battleground where several brands like Huawei, Nvidia, DeepSeek, and more are competing with each other. T...
Facts are intriguing, aren’t they? But, they also reflect a lot that give us a direction to our security strategy that needs to be solid to avoid the consequences of a zero-day attack. Even though it’s sound as a hypothesis, predi… #hackernews#news
The latest update for #Egnyte includes "A Faster, Smarter Way to Migrate From Microsoft SharePoint to Egnyte" and "6 Achievable Ways to Improve #DataGovernance".
NVIDIA AI Unveils ProRL Agent: A Decoupled Rollout-as-a-Service Infrastructure for Reinforcement Learning of Multi-Turn LLM Agents at Scale NVIDIA researchers introduced ProRL AGENT, a scalable inf...
A study reveals a novel framework with large language models to generate synthetic clinical trial data, addressing privacy issues and data scarcity. This method improves machine learning model performance in predicting trial outcomes, advancing clinical research.
📰 Custom Claude Skills: Automate Workflows Across Mac, iOS & Web (2026 Guide)
Anthropic’s new Claude Skills allow users to create and store reusable AI instructions for repeatable tasks, while Dispatch and Cowork extensions enable remote automation and Mac control. This ...
AI+DC Summit: Copyright protection in the AI era will be up to the courts, industry leaders say: At an Axios expert voices roundtable in Washington, D.C., on March 25, industry leaders and legal experts indicated that the complex issues surrounding copyright protectio… https://ranked.news/417093?u=b
Claude Code now has a /dream command that consolidates its memory between sessions, fixes stale dates, removes contradictions. Anthropic literally gave their coding agent REM sleep
DeepSeek AI: Everything You Need to Know (Full Guide 2026) The Chinese AI that shocked Silicon Valley, crashed Nvidia’s stock by $589 billion, and gave the world a free model that goes head-to-he...
Widely used Trivy scanner compromised in ongoing supply-chain attack
The widely used Trivy security scanner has been compromised in an active supply-chain attack. Administrators are advised to rotate secrets and credentials as a precautionary measure this weekend.
Why OpenAI is Losing $200 Million a Month and What it Means for ChatGPT OpenAI, once celebrated as a trailblazer in artificial intelligence with the unprecedented success of ChatGPT, now finds itse...
European Commission Confirms Cyberattack After AWS Account Breach The European Commission has confirmed a cyberattack on its cloud-based infrastructure following the identification of a compromised...
$400k for an AI-found bug in Solana? anatomist security got paid, sure. but i bet a human *told* the ai where to look. still impressive, but let's be real 🙄
[Sky News, VIDEO] A new study from Anthropic and the University of Toronto analysed 1.5 million conversations with the AI chatbot Claude, revealing rare but concerning cases of what some are calling "AI psychosis" #ai#realitydistortion#aipsychosis
(1/2) Anthropic is testing Mythos (Capybara), a "step change" above Opus. Leaked drafts warn it's "far ahead of any other model in cyber capabilities" and presages a wave of models exploiting vulnerabilities that "far outpace defenders." Thats some exciting stuff!
This issue is just going to get bigger and bigger. It reminds me of when email was first invented, and there were no standards built in to verify the sender or the content. We've been dealing with spam, phishing and malware ever since.
This week's Stellar Work newsletter: OpenAI killed Sora 30 min after a Disney planning meeting. A jury called Instagram a defective product. A Dutch court caught Grok lying about AI nudes. open.substack.com/pub/stellarw...
Google Gemini App March Update Adds AI Chat History Transfer Feature The post Google Gemini App March Update Adds AI Chat History Transfer Feature appeared on BitcoinEthereumNews.com . ...
Authenticated, But Not Verified: The Workforce Identity Gap
At RSAC 2026, Nametag CEO Aaron Painter sat down with Security Weekly to discuss a growing problem in enterprise security: organizations can no longer be certain who is actually behind a human or AI identity. #hackernews#news
ComfyUI 20% Faster on Linux! Can Pop!_OS Replace Windows?
- Linux has fast random access
- Especially faster with heavy workflows
- Operation feels similar to Android www.ai-image-journey.com/2026/03/linu...
"The last big story right before the war in Iran started was the collapse in the relationship between the Pentagon and Anthropic, with the latter objecting to any potential use of its models in either fully autonomous weapons or domestic surveillance. Of course, this story…"
LiteLLM Has a Free API — Unified Proxy for 100+ LLM Providers LiteLLM is a unified API gateway that lets you call 100+ LLM providers using the same OpenAI-compatible format. Switch between OpenAI...
Simon Brooke@simon-brooke.mastodon.scot.ap.brid.gy
"In 2018, more than 4,000 Google employees signed a letter opposing the company’s contract to build artificial intelligence for the Pentagon’s targeting systems. Workers organised a walk out. Engineers quit. And Google ultimately abandoned the contract."
Exclusive: Anthropic acknowledges testing new AI model representing ‘step change’ in capabilities, after accidental data leak reveals its existence is.gd/OA8GZp
1️⃣ A federal judge has temporarily blocked the Pentagon's ban on Anthropic, stating it was illegal retaliation for the company's First Amendment rights. The ruling highlights that the DOD failed to provide due process and did not follow legal procedures before implementing the ban.
Boston Metro Weather Forecast Bot@boston-4castbot.bsky.social
NWS Boston (via Facebook): Below normal high temperatures in the 30s for most, except in the lower 40s for portions of the CT River valley and southern RI. today. Dry weather continues, but with periods of clouds at time. [Link]#MaWX
How to Train Your Healthcare AI Agent: A Compliance-First Guide to Data Sharing
This guide breaks down how to build HIPAA-compliant AI agents in healthcare, covering BAAs, PHI handling, Safe Harbor de-identification, and FHIR-based data minimization. It explains why agentic AI ex… #hackernews#news
Fake Certificate Loader Conceals BlankGrabber Malware Chain BlankGrabber’s operators are experimenting with a stealthy loader chain that abuses Windows certificate tooling to hide a Rust‑based ...
www.reuters.com/technology/o... OpenAI going for stock listing and this year's strategy is to kill their darlings, i.e. any less profitable activity. Sora was much hyped video model, pulling plug on Disney partnership. Focus likely now to be coding and enterprise tools, but can they rival Anthropic?
I asked ChatGPT what to do after clicking a suspicious link — here’s what actually matters There's no need to freak out after you click on a suspicious link. Use these ChatGPT-generated tip...
A quantum reservoir autoencoder using reset noise channels suppresses shot-noise sensitivity by 10 orders of magnitude (MSE~10⁻¹⁴). A two-phase protocol enables blind decryption of unseen messages, outperforming variational circuit baselines; shared training data is the irredu...
5 Gemini sidebar tricks that made Chrome feel like a new browser I didn’t expect much from Gemini in Chrome until it became part of everything I do online. The Chrome browser started to feel like...
🚨 Using Claude.ai on your site? Check this.
A CLAUDE.md file in your web root can contain keys, creds & API secrets, and it’s publicly accessible at yoursite.com/CLAUDE.md.
✅ Move it outside the web root
✅ Use environment variables
✅ Add it to .gitignore
New AI = new attack vectors. #cybersecurity
Google just revised its ‘Q-Day’ timeline: Quantum computers could break existing encryption techniques within three years – and enterprises are nowhere near ready | Technical progress means Q-Day could arrive a lot sooner than expected, according to Google | ITPro
Yep, I got to a point where anything I said to anyone in person or on the platforms or digital access points of people who abused me would be spun so I hung out and learned the mechanisms behind the ai tools and data processing as a survivor with cptsd. #crimeanalytics
A look at some themes at this year's Hill and Valley Forum: embracing government-led industrial policy to onshore manufacturing, AI's unpopularity, and more (Newcomer)
OpenAI is throwing everything into building a fully automated researcher
OpenAI is redirecting its research efforts toward building a fully automated AI researcher system capable of independently tackling large, complex problems. The San Francisco company aims to create…
What is AI + Zero Trust? Why It is on Every Security Roadmap Now AI and Zero Trust together are quickly becoming the baseline to keep up with evolving attacks, hybrid environments, and rising compl...
autonomous AI agent traffic on the internet grew 7,851% in a single year. OpenAI bots alone account for 69% of all AI-driven web traffic. humans are now the minority on their own internet
A new AI review! lagonapp/lagon ⭐3.2/5.0
Lagon is an ambitious open-source edge/serverless platform that combines a Rust-based JavaScript runtime (V8 isolates) with a TypeScript-centric developer experience (CLI, dashboard, integrations, examples). https://gitrated.com/lagonapp/lagon
From CTFs to Real Investigations
AI & Media Verification
Geolocation & Environmental Analysis
OSINT Challenges & Structured Exercises
Real-World OSINT Operations
This one Linux shell character made me feel like a hacker Who know it would be so easy to create programs from the command line? Like many people using Unix-like operating systems for the first tim...
Google is dramatically shortening its readiness deadline for the arrival of Q Day, the point at which existing quantum computers can break public-key cryptography algorithms that secure decades’ worth of secrets belonging to militaries, banks, governments, and nearly every individual on earth […]
ML identifies the thermodynamic arrow of time on a 10-qubit NV-center diamond processor. A CNN achieves ~92% accuracy, k-means ~91%, and a diffusion model reproduces entropy production and directional heat flow without prior physical knowledge.
Remarkable fact from @dkthomp.bsky.social: "Between December 2025 and March 2026, the AI lab Anthropic more than doubled its annual recurring revenue from $9 billion to more than $20 billion. According to several analysts, there is no record of any company growing this fast at this scale...ever. "
Lockdown Mode is easy to turn on and off, and can protect your security in very specific situations. Here’s how it works and when to use it. ssd.eff.org/module/how-...
A.A. Al Moosa Enterprises (ARENCO Group), a Dubai-based conglomerate, has been targeted by a ransomware attack attributed to the threat actor payload. Incident discovered on March 28, 2026. #RansomwareAttack#DubaiBusiness#UnitedArabEmirates
Exclusive: Anthropic left details of an unreleased model, invite-only CEO retreat, sitting in an unsecured data trove in a significant security lapse fortune.com/2026/03/26/a...
When HSE Recruitment Process&Its Payroll are fully qualified to pay HSE Staff appropriately from date started,correct location,right staff number&correct terms,employment law,PRSI, PAYE,
It has been proven that our voting machines have been compromised. We need to take action ASAP, before the '26 election is also compromised. cmarmitage.substack.com/p/in-7-month...
A new AI review! alacritty/alacritty ⭐4.1/5.0
Alacritty is a mature, high-performance, cross-platform terminal emulator written in Rust with a clear focus: render fast, integrate with existing tools, and stay configurable without becoming a full terminal “... https://gitrated.com/alacritty/alacritty
Critical flaws in PTC Windchill/FlexPLM exploited with mitigations but no patch; Langflow CVE-2026-33017 active for RCE, upgrade to 1.9.0. Ransomware targets energy firms; Nova Scotia Power breached. #Germany#Ransomware#Canada
#Google is rolling first broad #coreupdate in 2026, it's business as usual with some novelties, including the usage of #Gemini 4.0 Semantic filter that should detect and demote AI generated content.
NanoClaw + Docker just showed up in Forbes’ CIO newsletter.
The conversation is shifting - not just what agents can do, but how to run them safely once they’re doing real work.
Isolation isn’t just a feature anymore. It’s becoming the starting point.
“There is also a significant cybersecurity risk,” Tordjman added, “if hackers were to gain access to a hospital’s network and inject synthetic images to manipulate patient diagnoses or cause widespread clinical chaos.”
🫥😬
Focus on Local? Let's call this the "Near Me ChatGPT Update" :) -> It's a big update by ChatGPT IMO that rolled out on March 26th that nobody is talking about... You can now share your device location so ChatGPT can provide more relevant information like local recommendatiaons, news, and weather.
From @edweek.org: Hackers are using generative AI to launch sophisticated attacks on K-12 schools. But experts say the fight is far from lost — there are practical, low-cost steps every district can take right now to strengthen their defenses.
OpenAI sets two-stage Sora shutdown with app closing April 2026 and API following in September
OpenAI announced it will shut down Sora, its AI video generation tool, in two phases: the app will close in April 2026 and the API will follow in September 2026. The decision…
OpenAI just canceled its video-generating app Sora and put the brakes on AI-generated videos in ChatGPT. Guess the hype was too much to handle {kcurry.bsky.social/p/123456}
💻 Anthropic's powerful new Claude Mythos AI model was accidentally leaked, showing major advances in reasoning and coding but raising concerns about cybersecurity risks.[1][2]
Budget constraints and compliance considerations have presented hurdles for the agency as it aims to onboard technologists, a DOE IT leader said during a Thursday event. via @lindseywilkinson.bsky.socialfedscoop.com/energy-depar...
Like discovering that Iran could block the straight of Hormuz, last week Hegseth found out that the First Amendment blocked his "supply chain risk" designation of Anthropic.
A new AI review! FuelLabs/fuel-core ⭐4.1/5.0
`fuel-core` is a large, actively maintained Rust workspace implementing the Fuel node/client, exposing a GraphQL API and bundling the operational tooling needed to run local networks, devnets/testnets, and produc... https://gitrated.com/FuelLabs/fuel-core
There's a new cycle developing in some markets - one tied to the emergence of rumours to what the Next Big Thing will do. IN this case, rumours of what Anthropic's upcoming Mythos model will be able to do - AI to defeat other peoples products eg cybsecurity. www.cnbc.com/2026/03/27/a...
"Iranian government hackers are using Telegram as a way to steal data from hacked dissidents, opposition groups, and journalists who oppose the regime around the world, according to an FBI alert" techcrunch.com/2026/03/23/f...
📰 Claude Paid Subscriptions More Than Double in 2026 Amid Consumer Surge
Claude paid subscriptions have more than doubled this year, signaling explosive growth in consumer adoption. Though Anthropic hasn't disclosed exact user numbers, industry estimates place total user...
anthropic shipped 14 products in march, had 5 outages, and accidentally leaked their most powerful model. MCP hit 97 million downloads, Claude Code hit $2.5B ARR, and the servers could not take it
The letter explained that the Immigration Department rejected Kémy Adé's permanent residence application because the duties listed for her job did not match the Canadian work experience she claimed: Canada used AI to review immigration—then this happened
Fully homomorphic encryption allows you to get information without revealing your personal data, but it’s currently a slow, intensive process. But #Intel’s new Heracles chip speeds up FHE computing 5,000-fold compared to a CPU. spectrum.ieee.org/fhe-intel
A look at the decadelong feud between Sam Altman and Dario Amodei; sources say Amodei likened Altman's legal fight with Musk to Hitler's fight with Stalin (Keach Hagey/Wall Street Journal)
Google is rolling out a new feature called “Results about you” to help users scrub more sensitive information such as social security and driver’s license numbers from internet se…
Google announced Search Live has launched globally, for all languages and locations where AI Mode is available. Google said this is possible by its "new audio and voice model, Gemini 3.1 Flash Live, which delivers even more natural and intuitive conversations."
Do you worry about which apps are watching you? A new Android update will now show a clear "blue dot" whenever an app is accessing your location. It's a simple way to stay in control of your privacy without being a tech expert!
