Content Buffer
Welcome, Tech Leaders.
The NSA has quietly started using Mythos Preview from Anthropic, a frontier model announced earlier this month for cybersecurity tasks. The Department of Defense had previously labeled Anthropic a 'supply chain risk' due to its refusal to grant unrestricted access to the model's full capabilities. This move highlights the growing reliance on AI in national security and raises concerns about the potential risks involved.
WhatsApp is testing a new subscription tier called WhatsApp Plus, which offers a range of customization options. Users can expect features like expanded pinned chats, custom lists, new chat themes, and more. The company hasn't specified pricing for the paid plan, but WABetaInfo suggests it might be priced at €2.49 per month in Europe or 229 PKR ($0.82) in Pakistan. A one-month free trial is being offered to users. This move could impact WhatsApp's business model, which now focuses on allowing companies to reach users and create click-to-WhatsApp ads. Let's dive deeper…
In today's ContentBuffer update:
Deezer's Music Library Overwhelmed by AI-Generated Tracks
Vercel's internal systems breached, customer data accessed
Anthropic Flags Claude Mythos 5 Under ASL-4, Holds Back Public Release
NVIDIA and Apple Bring CloudXR 6.0 Natively to Apple Vision Pro
5 new AI tools & 5 new AI jobs
More tech news
Latest Development
Ai-security
Image source: techcrunch.com
Summary: The NSA is using Anthropic's Mythos Preview for top-secret cybersecurity tasks, sparking questions about the US military's expanding use of AI tools.
Key Points:
40 organizations have limited access to Mythos, including the UK's AI Security Institute and undisclosed recipients like the NSA, which has been using it for several weeks.
Mythos was designed specifically for cybersecurity tasks and boasts a benchmark score of 95% accuracy in detecting zero-day attacks.
The U.S. military is expanding its use of Anthropic's tools, with a recent meeting between Anthropic's CEO and White House officials seen as productive.
Anthropic refused to allow Pentagon officials unrestricted access to its model's full capabilities, citing concerns over mass domestic surveillance and autonomous weapons development.
The Department of Defense labeled Anthropic a 'supply chain risk' in 2022, but the company has since thawed out its relationship with the Trump administration.
Why it matters: If you're working on sensitive projects or relying on AI-powered security tools, this development may impact your workflow and decision-making. The NSA's use of Mythos Preview underscores the growing importance of AI in national security and highlights potential risks involved.
Messaging-platforms
Image source: techcrunch.com
Summary: WhatsApp tests new paid tier with customization options; pricing unclear
Key Points:
1. WhatsApp Plus is priced at €2.49 per month for users in Europe.
2. In Pakistan, the premium tier costs 229 PKR ($0.82), a significant markup from the standard service.
3. The test group will have access to a one-month free trial before committing to the paid plan.
4. WhatsApp Plus offers up to 20 pinned chats, a 6-fold increase from the current limit of three.
5. Custom themes, ringtones, and notification tones are available for users who opt-in to the premium tier.
Why it matters: If you're running a business on WhatsApp, this new paid tier could impact your customer engagement strategy. With customization options, businesses may need to rethink their approach to interacting with customers. This move also highlights the shift in WhatsApp's business model from paid messaging to ad-based revenue.
Music-generation
Image source: techcrunch.com
Summary: Deezer's overwhelmed with AI-generated tracks, making up 44% of new uploads. The company receives over 75K fake tracks daily, but most go unnoticed. This raises questions about music authenticity and the role of AI in content creation.
Key Points:
1. In January 2025, Deezer received around 10,000 AI tracks per day, a significant drop from 60,000 in the same month last year.
2. The company receives more than two million AI-generated tracks per month, with almost 75,000 arriving daily, according to recent data.
3. Consumption of AI-generated music on Deezer is still low, at just 1-3% of total streams, despite the massive influx of AI tracks.
4. A staggering 85% of AI-generated stream detections are fraudulent and demonetized by Deezer, highlighting the platform's ongoing struggle with AI-music abuse.
5. In a recent survey, 97% of participants couldn't tell the difference between fully AI-generated music and human-made music, raising questions about the value of AI tracks.
Why it matters: If you're a developer working with audio processing or music streaming services, this raises questions about the authenticity and ownership of digital content. As AI-generated tracks become more prevalent, how will platforms like Deezer ensure transparency and fairness for creators?
Cybersecurity
Image source: techcrunch.com
Summary: Vercel's internal systems breached by Context AI app; hundreds of users may be affected
Key Points:
Hackers exploited a Context AI app to access Vercel's internal systems and customer data.
The breach originated from an employee's corporate account hosted on Google, which was connected to the Context AI app via OAuth.
Some of Vercel's internal systems weren't encrypted, making it easier for the hackers to exploit.
Next.js and Turbopack projects were not affected by the breach.
Vercel has contacted customers whose data and keys were compromised.
Why it matters: If you're running Next.js or Turbopack on Vercel and haven't reviewed your app's security settings recently, this breach is a wake-up call. The hackers accessed sensitive customer data, including API keys and source code, which could lead to downstream breaches if not addressed promptly. This incident highlights the importance of secure workflows, especially for teams using cloud-based services like Vercel.
Ai Models
Summary: Anthropic confirmed Claude Mythos 5 in early April 2026 but said the model will not be released publicly or via standard API because internal testing…
Key Points:
Model classified at AI Safety Level 4 (ASL-4)
No public or standard API release
First widely disclosed ASL-4 classification from a frontier lab
Why it matters: It is the first public example of a frontier lab withholding a top model on safety grounds under a formal responsible-scaling policy — a milestone moment for AI governance.
Hardware
Image source: nvidianews.nvidia.com
Summary: NVIDIA and Apple announced a collaboration bringing native integration of NVIDIA CloudXR 6.0 to visionOS, delivering NVIDIA RTX-powered simulators an…
Key Points:
Native CloudXR 6.0 integration on visionOS
RTX-powered simulators and pro 3D apps on Vision Pro
First major NVIDIA-Apple co-announcement in recent memory
Why it matters: The tie-up gives Apple Vision Pro a credible enterprise XR content pipeline and positions Apple and NVIDIA as co-leaders of pro-grade spatial computing.
New Tools & Job
AuraWrite AI - Humanize AI text from ChatGPT, Claude, Gemini, and more with AuraWrite's AI Humanizer. Make your content more natural, engaging, & human in seconds. Try now!
Genaraera - Create professional infographics with AI in seconds. Infographics AI offers the best infographics generator for digital marketing, data visualization, and content creation. Try our AI-powered infographics tool free today.
Nano Banana Pro Studio - Create professional 4K images in 3-10 seconds with Gemini 3 Pro Image. Advanced multimodal AI delivers studio-grade visuals and precise text control.
UX Magic - Create stunning AI-generated mockups and figma ready wireframes in seconds with UXMagic. Simplify UX design workflows and boost creativity—no design skills needed.
Rizzy Agent - Rizzy is your 24/7 AI Lead Generator on X (Twitter). It finds people talking about what you sell and sends them straight to you.
Product Engineer - We are hiring exceptional product engineers who can fully own projects and get them done.
Senior Technical Lead - We're hiring our founding technical hire — a Senior Technical Lead who will own the entire technical vision and execution.
UI Engineer - Build our UI from the ground up, establishing clear, maintainable patterns for components, state management, styling, testing, performance, and accessibility.
Senior DevOps Infrastructure Engineer - We are seeking an experienced Senior DevOps Infrastructure Engineer to architect, build, and operationalize a modern cloud-native delivery platform on AWS from scratch.
Senior Product Security Engineer - Drive critical product security initiatives across Vercel's products and platform.
QUICK HITS
Adobe Summit 2026 Opens in Las Vegas With AI Creative Workflows Center Stage - Adobe Summit 2026 kicks off April 19 and runs through April 22 in Las Vegas, with the agenda dominated by generative AI, creative workflows, and cust…
NVIDIA Posts 73% Sales Jump to $68B in Q4 FY2026 - NVIDIA reported Q4 FY2026 revenue of $68 billion, up 73% year-over-year, with non-GAAP EPS climbing 82% to $1.62. The company expects revenue to acce…
Fortune: AI Holdouts Are Running Out of Time as Adoption Data Shifts - A Fortune analysis published April 16 argues that worker and executive resistance to AI is losing ground as productivity, hiring, and compensation da…
Tesla AI5 Chip Tapes Out With Samsung and TSMC as Manufacturing Partners - Tesla's AI5 inference chip was officially taped out on April 15, 2026, with Samsung and TSMC handling fabrication. Tesla claims a single AI5 matches …
ChatGPT Suffers Major Global Outage, Taking Down Conversations, Login and Voice Mode - OpenAI's ChatGPT went down for thousands of users globally on April 20, with a 'partial outage' lasting at least 90 minutes starting around 10:05am E…
Tesla Activates Fully Unsupervised Robotaxis in Dallas and Houston - On April 18, 2026, Tesla turned on fully unsupervised robotaxi rides in Dallas and Houston with no safety monitors in the vehicles. The rollout uses …
THAT'S A WRAP
Partner with Us
Showcase your product to our engaged community of Tech enthusiasts
Our newsletter reaches a diverse audience of tech professionals, investors, engineers, managers, and business leaders worldwide. Get in touch today.
Comments
Be the first to comment
Want this in your inbox?
Get the daily edition delivered free every morning.
Latest Issues
View Archive →
Elon Musk's legal battle against OpenAI CEO Sam Altman entered a new phase as jury selection concluded, despite some jurors harboring negative views of Musk. The case centers on alleged broken promises at OpenAI.
On-demand home services startup Snabbit has closed a $56 million Series D round, valuing the company at around $350 million. The funding will fuel growth as demand surges.
Weiyao Wang, a top Meta AI expert, has left the company for Thinking Machines Lab (TML), which just signed a multibillion-dollar cloud deal with Google. This is a major blow to Meta's AI efforts.
Meta just signed a massive deal with AWS, dumping its previous $10 billion agreement with Google Cloud. The move will see Meta use millions of ARM-based Graviton chips for AI-related compute needs.
Sean Plankey has withdrawn his nomination to lead the US federal cybersecurity agency CISA due to a Senate holdup. Sen. Rick Scott is blocking Plankey's appointment over a Coast Guard contract unrelated to cybersecurity.
Meta's rolling out a unified account system, simplifying user experience across its ecosystem. With 200M users affected, here's how it'll change your login process.