NerdNewsJune 05, 2026 |
|
News & Trends
Mythos Outperforms GPT5.5 on Google Chrome Vulnerability Exploits
Anthropic's Claude Mythos outperformed OpenAI's GPT5.5 in a new benchmark, ExploitBench, designed to test AI models' ability to exploit real-world vulnerabilities. Mythos achieved a higher exploitation performance, reaching the highest tier on 21 of 41 vulnerabilities, while GPT-5.5 reached the top tier on just two cases. This highlights the potential of AI models in accelerating vulnerability discovery and exploitation, but experts urge caution and emphasize the need for defenders to develop AI-driven remediation at scale.
Meta's Unreleased Facial Recognition Feature
Wired discovered code for an unreleased facial recognition feature, called 'NameTag', in Meta's AI app. The feature can capture faces using smart glasses and notify the wearer when it recognizes a previously captured face. Meta has not announced any plans to release this feature, but it has explored facial recognition technology in the past, including its use in Facebook and Instagram.
Apple Approves Poke as First AI Agent on Messages for Business
Poke, a startup, becomes the first AI agent approved to run on Apple's Messages for Business platform, allowing businesses to communicate with customers through iMessage. Poke can help with daily planning, calendar management, and more via text message.
Meta Rolls Out AI Creator Assistant on Facebook
Meta has launched a new AI creator assistant on Facebook, providing personalized recommendations and performance insights to creators. The assistant can help with content ideas, audience engagement, and language translations, aiming to keep creators active on the platform.
WWDC 2026: Siri's Revamp and Apple Intelligence Updates
Apple's WWDC 2026 is expected to unveil a major AI upgrade to Siri, transforming it into a more conversational assistant. Other updates include an AI agent app store, Camera and Photos app enhancements, and Apple Wallet updates. |
|
Options & Tutorials
AI Weather Startup Out-Forecasting Government Agencies
WindBorne Systems' WeatherMesh-6 AI model provides more frequent and accurate weather forecasts than traditional models, including those from the European Centre for Medium-Range Weather Forecasts. The model uses data from 400 balloons and improves upon traditional models by ingesting data directly, allowing for more accurate predictions.
Publishers can now opt out of Google AI summaries and training
The UK's competition watchdog has ruled that Google must provide online publishers with the ability to opt out of their work being summarized by artificial intelligence or used to train Google's models. This change gives publishers more control over how their content is used and allows them to prevent their work from being used in Google's AI features.
Cisco Live 26: Networks the key in post-Mythos world
Cisco has introduced Cloud Control, a platform that brings together humans and AI agents to manage critical IT infrastructure. The platform offers a single management plane, cross-domain telemetry, and purpose-built models, allowing for system intelligence that scales with complexity. Cisco also announced other products, including AI Canvas and Cloud Control Studio, to support the platform.
Spiral 4.0 Goes Agent-Native
Spiral 4.0 introduces agent-native access via MCP, CLI, and API, team workspaces, and a style engine that generates writing indistinguishable from the user's own 87% of the time. The update also includes a $10 price drop, with personal plans starting at $15/month.
Top Secrets Management Tools
This article reviews 16 secrets management tools, including enterprise, cloud-native, open-source, and Kubernetes/workload identity tools. It covers key features, ideal use cases, and decision frameworks to help choose the best tool. |
|
Launches & Tools
Samsung Updates Health App with Gen AI
Samsung's updated Health app uses generative AI to analyze biometric data, providing simple and actionable guidance. New features include Vitals, Heart Health Score, Daily Cardio Load, and Fitness Index, which offer personalized advice and recommendations for improvement. The app will be fully realized with the launch of Samsung's next generation of Galaxy Watches.
Anthropic's IPO Plans
Anthropic's co-founder Daniela Amodei discusses the company's plans for an IPO, citing the need for capital to train and serve AI models. Despite concerns about AI returns, Amodei believes businesses are still early in figuring out how to deploy AI effectively.
Ramp raises $750M at $44B valuation
Ramp, a corporate expense management platform, has raised $750 million at a valuation of $44 billion. The company has expanded its services to include payments, fraud detection, and accounting, and has built an AI story around itself. Ramp's annualized revenue is over $1 billion and it has over 70,000 customers.
New IronWorm malware hits 36 packages in npm supply-chain attack
A new supply-chain attack has infected 36 packages on the Node Package Manager (npm) index with IronWorm malware, which targets environment variables and credential files, and can self-propagate using stolen credentials.
Windows is back on the Microsoft menu
Microsoft focused on Windows at its Build 2026 conference, highlighting its importance in the company's AI efforts. The company unveiled new Surface hardware, including the Surface Laptop Ultra, and emphasized the potential of local AI compute with Nvidia's RTX Spark chips. Microsoft also discussed its plans for hybrid compute, where local devices handle some workloads and the cloud handles others. |
|
Quick Links
Raise Security Concerns with Procurement Now, Because Quantum Can’t Wait
Security leaders must accelerate transition plans to post-quantum cryptography through proactive inventorying, procurement, and crypto-agility efforts. Only 8% of SSH servers worldwide currently support post-quantum cryptography, and the countdown to cryptographically relevant quantum computers has already begun.
Canada Prime Minister Mark Carney announces national AI strategy
Canada's Prime Minister Mark Carney has announced a new national AI strategy called 'AI for All', which aims to strengthen data protections, increase AI adoption, and invest in AI infrastructure. The plan includes updating legislative frameworks, creating an online safety regime, and providing free AI training. However, the strategy has been met with skepticism due to concerns about the impact of AI on productivity and its growing unpopularity.
Meta builds data centers in tents
Meta has built six 'rapid deployment structures' or tents outside of New Albany, Ohio, to house its multi-gigawatt data centers, cutting construction time in half. The strategy is reminiscent of Tesla's use of tents in its factory parking lot. The tents will be powered by 200 megawatts of modular gas turbines and will house AI chips worth billions of dollars.
Helion Raises $465M
Helion, a fusion startup backed by Sam Altman, has raised $465 million to build a power plant for Microsoft, with a goal to deploy fusion power to the grid as early as 2028. The company's approach to fusion power uses magnets to compress fuel and harvest electricity directly from the magnets, which could improve efficiency but has raised skepticism among some experts.
Waymo's Spent Robotaxi Batteries to be Used as Grid Storage
Waymo has announced a deal with energy storage company B2U to use retired robotaxi batteries for grid storage in California and Texas, providing a solution for the end-of-life treatment of thousands of robotaxis and promoting sustainable energy practices. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |