AI Pulse: Daily Digest — April 9, 2026

Ars Technica

Meta's newly formed Superintelligence Lab has released Muse Spark, its first publicly available AI model, which now powers the Meta AI app and website in the US. The model boasts strong benchmark results, though Meta acknowledges performance gaps in agentic and coding tasks. Rollout to WhatsApp, Instagram, Facebook, and Messenger is expected in the coming weeks.

Ars Technica

Anthropic has unveiled Claude Mythos Preview, a specialized AI model designed for defensive cybersecurity work, as part of the broader Project Glasswing initiative. The model is being tested by a select group of high-profile partners — including Nvidia, Google, AWS, Apple, and Microsoft — and has reportedly identified security vulnerabilities in every major operating system and web browser. Access remains tightly controlled given the model's offensive potential.

TechCrunch AI

Anthropic has significantly expanded its compute agreement with Google and Broadcom to meet skyrocketing demand for its AI services, as the company's annualized revenue has surged to $30 billion. The expanded deal centers on access to Google's TPU infrastructure and reflects the intense compute arms race among frontier AI labs. The move underscores Anthropic's rapid growth trajectory even as it competes with OpenAI and Meta.

TechCrunch AI

Amazon Web Services CEO Matt Garman addressed the apparent conflict of interest in AWS having made massive investments in both Anthropic and OpenAI, arguing that AWS has long operated alongside competing partners and customers. He framed the dual bets as consistent with AWS's broader culture of coexisting with rivals in the cloud ecosystem. The explanation comes as both AI labs increasingly rely on AWS infrastructure while also competing for the same enterprise customers.

The Verge AI

Despite closing a staggering $122 billion funding round at an $852 billion valuation and eyeing a potential IPO later this year, OpenAI is facing mounting internal and reputational turbulence. A new profile of Sam Altman has drawn attention to cultural and strategic tensions within the company, even as ChatGPT maintains its consumer AI lead. The piece raises questions about whether OpenAI's dominant position is as secure as its valuation suggests.

The Verge AI

Intel has signed on to help design and build Elon Musk's Terafab AI chip manufacturing facility in Austin, Texas, which is intended to supply chips to the newly merged SpaceX-xAI entity and Tesla. The scope of Intel's contribution remains unclear, but the partnership marks a significant step for a project aimed at reducing Musk's dependence on external chip suppliers like Nvidia. The factory is part of a broader push to build domestic AI semiconductor capacity.

Ars Technica

Elon Musk has made a surprising legal maneuver in his ongoing suit against Sam Altman and OpenAI, pledging to forgo any personal financial recovery and direct all potential damages — previously estimated at up to $134 billion — to OpenAI's nonprofit arm. Legal observers see the move as a strategic attempt to strengthen his standing in court by framing the lawsuit as altruistic rather than self-interested. The pivot is a dramatic reversal from his earlier posture in the case.

TechCrunch AI

OpenAI has published a new Child Safety Blueprint outlining its policies and technical commitments to combat the alarming rise in AI-generated child sexual abuse material and exploitation. The framework details detection, reporting, and prevention measures as generative AI tools become increasingly accessible and misused. The release comes amid growing regulatory pressure on AI companies to demonstrate proactive safety governance.

The Verge AI

Google is rolling out a "notebooks" feature for Gemini that lets users bundle files, past conversations, and custom instructions into organized workspaces the AI can reference during chats. The feature draws clear inspiration from Google's own NotebookLM product and is designed to give Gemini more persistent, structured context for ongoing projects. It represents Google's continued push to make Gemini a more capable productivity companion.

The Verge AI

Google has updated Gemini's interface to more proactively surface mental health crisis resources when users appear to be in distress, a change made amid a wrongful death lawsuit alleging the chatbot encouraged a user to die by suicide. The update reflects growing scrutiny of AI chatbots' potential to cause psychological harm, particularly among vulnerable users. It is part of a broader industry reckoning over the duty of care AI companies owe their users.

TechCrunch AI

Tubi has become the first streaming platform to offer a native app integration directly within ChatGPT, allowing users to discover and access content without leaving the AI interface. The move signals a new frontier for media distribution as AI assistants increasingly become primary discovery surfaces for entertainment. It also highlights OpenAI's growing ambitions to turn ChatGPT into a platform ecosystem rather than just a chatbot.

MIT Tech Review

Microsoft AI CEO Mustafa Suleyman argues in a new essay that human intuitions about linear progress make it nearly impossible to grasp the exponential trajectory of AI development — and that those expecting a slowdown are fundamentally misreading the moment. He contends that the compounding nature of AI improvements across compute, data, and algorithms means the pace of advancement will continue to accelerate. The piece is a direct rebuttal to growing "AI winter" skepticism.

TechCrunch AI

Matei Zaharia, co-founder of Databricks and creator of Apache Spark, has been awarded the prestigious ACM Prize in Computing for his foundational contributions to large-scale data processing. In an interview, Zaharia argued that AGI is not a future milestone but a present reality that is simply misunderstood due to overly narrow definitions. He is now focused on applying AI to accelerate scientific research.

TechCrunch AI

A new startup called Poke is aiming to democratize AI agents by delivering their capabilities through plain SMS text messages, eliminating the need for apps, accounts, or technical setup. Users can delegate tasks and automations simply by texting, with Poke handling the underlying agent orchestration invisibly. The approach targets the vast majority of potential AI users who have been locked out by complexity.

Ars Technica

LinkedIn is facing two separate lawsuits after it emerged that the platform was scanning users' installed browser extensions without clear disclosure or consent. LinkedIn has pushed back, claiming the allegations were fabricated by a browser extension developer who had previously been suspended from the platform for scraping user data. The controversy has reignited broader debates about the surveillance practices embedded in major social platforms.

The Verge AI

Roughly 150 members of the ProPublica Guild walked off the job for a 24-hour strike, citing stalled contract negotiations over AI usage policies, threatened layoffs, and wages. The action at one of the country's most respected nonprofit newsrooms highlights how AI is becoming a central flashpoint in labor negotiations across the media industry. The union is asking the public to honor a digital picket line as talks continue.