Edition: July 7, 2026

Meta has launched Muse Image, a new AI image generator that allows users to create and edit images, but the feature is already facing backlash over its use of public Instagram photos without explicit consent. The model can manipulate another user's images with AI, sparking concerns over privacy. Meta claims users have control over the feature, but critics argue it's a 'privacy landmine' waiting to detonate.

Meta has introduced a new AI image model that allows users to generate images using public Instagram photos, unless the account owner opts out. The feature is integrated into the Instagram app and can be disabled by toggling the 'Allow people to use your content on Instagram and with AI features on Meta' setting. Users will not be notified when their content is used to generate AI images.

Hugging Face and Amazon SageMaker Studio have announced a deep-link integration, allowing developers to go from model discovery to hands-on experimentation in SageMaker Studio with a single click. This integration streamlines the path from inspiration to experimentation, reducing friction and enabling faster iteration. The integration introduces three new capabilities, including deep links, pre-configured permissions, and GPU quota visibility.

Meta has launched its new Muse Image model, an AI image generation tool that can pull other Instagram users into AI-generated photos, and will be integrated across Meta's platforms. The model is part of the Muse family of AI models and is described as 'agentic', working with a large language model to reason through prompts. It will be available on Instagram, WhatsApp, and soon on Facebook and Messenger.

A new command line tool called Atrophy aims to help coders reverse the decline in their programming skills caused by relying on AI agents. The tool provides regular drills in five skill areas, including syntax recall and debugging, to reinforce learning. By using Atrophy, developers can identify and address skill gaps before they become a major issue.

Discord has admitted to a bug in its AI moderation system that wrongfully banned over 8,000 users due to harmless images being flagged as harmful content. The issue, which was identified and fixed, highlights the challenges of AI-assisted moderation. All affected accounts are being restored.

Anthropic's Claude Cowork, a general knowledge work agent, is now available on web and mobile for Max subscribers, allowing users to start tasks on one device and pick up where they left off on another. This expansion aims to make Cowork feel like an administrative coworker, working in the background across devices. The move signals a shift in AI firms' focus from chatbots to integrating their products into everyday work surfaces.

Anthropic has announced that its AI agent Claude Cowork can now run on smartphones and web browsers without requiring a desktop connection, allowing users to access its task automation features on-the-go. This update eliminates the need to keep laptops open to keep the agent running, making it more convenient for users. The feature is initially available as a beta to subscribers of Anthropic's Max plan.

Microsoft has announced the integration of Hugging Face models on Foundry Managed Compute, a managed GPU platform-as-a-service, allowing for easy deployment of open-source models with enterprise security and governance. This integration provides developers with a curated catalog of open-weight models from the Hugging Face ecosystem, refreshed weekly. The platform offers a range of features, including content safety filters, task-adherence guardrails, and unified RBAC.

Insilico Medicine is advancing an AI-identified drug, rentosertib, to Phase III trials for treating idiopathic pulmonary fibrosis (IPF), a disease with a median survival rate of two to four years post-diagnosis. The drug has shown promising results in a randomized trial, with patients demonstrating a mean forced vital capacity gain. This progression marks a significant step in the computational drug discovery sector.

OpenAI CEO Sam Altman proposes giving the US government a 5% stake in the company, potentially valuing at $320 per American household, to address concerns over AI's impact on labor markets. A leaked Treasury report compares the AI market to the dotcom bubble, contradicting the administration's public optimism. Samsung's profits have jumped 1,800% due to booming AI chip sales, but shares slumped over fears of a stalled AI boom.

A new security startup, Savi Security, has launched an app to protect consumers from realistic AI-generated scams, such as kidnappers demanding ransom, after the founders' mother was targeted in a similar scam. The company has raised $7 million in seed funding and is launching its app for iPhone and Android. The app aims to provide real-time intervention to detect and prevent such scams.

IT leaders can make informed decisions about AI investments by focusing on foundational elements of AI architecture, including data preparation and context engineering. A stable AI architecture allows for reliable and integrated systems at scale. Effective AI strategy begins with connecting and organizing data across the organization.

A viral video of Norwegian striker Erling Haaland was revealed to be an AI-generated deepfake, yet it continued to spread online as fans have become accustomed to consuming athletes as characters with recognizable quirks and storylines. This phenomenon is part of a larger shift in sports fandom, where athletes are treated like fictional characters and fans actively participate in creating content. The ease with which the deepfake was accepted highlights a growing trend where fans prioritize entertainment over reality.

L'Oreal, Mondelez, and Nestle are leveraging AI to accelerate product development, using the technology to predict ingredient performance, generate recipe ideas, and optimize formulations. This approach has enabled L'Oreal to make product formulation four times faster and Mondelez to reduce the number of physical samples needed during development. The use of AI is part of a broader innovation push in the consumer goods industry.

A British space startup, Mass Balance, has launched a longevity lab into orbit to study disease-causing proteins in zero-gravity conditions, which could lead to breakthroughs in age-related diseases. The autonomous laboratory will orbit the Earth for a couple of months, collecting data on how live cells grow and react under weak gravity. This research could potentially fill gaps in training data for life sciences models and aid in the development of new medicines.

US-based Forterra has deployed over 100 autonomous ground vehicles in Ukraine, marking the largest deployment of its kind in combat by a US defense tech company. The vehicles have driven over 2,500 miles and completed 52 casualty evacuations since their arrival in October. The deployment highlights the growing importance of autonomous vehicles in modern warfare.

Google DeepMind has expanded its Managed Agents in Gemini API, introducing new capabilities such as background execution, remote MCP server integration, custom function calling, and credential refresh. These updates aim to help developers build reliable, production-ready agents. The new features enable managed agents to operate as asynchronous workers in real development environments.

Hugging Face and SkyPilot have partnered to enable running AI workloads on any cloud while storing data on Hugging Face, eliminating egress fees. This integration allows for seamless access to models and datasets across different cloud platforms. The partnership enables zero-egress storage, reducing costs and increasing efficiency for AI development and deployment.

LeRobot v0.6.0 introduces new features such as world model policies, reward models, and simulation benchmarks to improve robot learning. The update includes six new simulation benchmarks, a deployment CLI, and cloud training on HF Jobs. It also brings depth sensing, custom video encoding, and a leaner installation.

The rise of open source AI models is not currently hurting companies like Anthropic, as they continue to dominate the market for premium AI models, with open source models serving as a cheaper alternative for more mature use cases. Despite the growth of open source models, Anthropic still accounts for over half of the overall AI spend on some platforms. The AI economy may be developing a two-tiered structure, with frontier labs owning discovery and open source models owning production.

OpenAI's chief futurist, Joshua Achiam, is leaving the company after nearly nine years, citing a desire to work on the company's mission from outside its walls. Achiam's departure is the latest in a series of exits from the company's safety-focused leadership team. His role, which involved studying the potential harms and benefits of artificial intelligence, may not be filled immediately.

Microsoft is reducing its reliance on third-party AI models from OpenAI and Anthropic, instead using its own in-house MAI models to cut costs in programs like Excel and Word. This move is part of a broader trend in the tech industry to curb spending on AI services. Microsoft's decision reflects the increasing cost of AI and the need for companies to find more affordable solutions.

Anthropic's Claude Cowork AI platform is now available on mobile and web, expanding access beyond the desktop app. The rollout starts with Max subscribers, with other plans to follow in the coming weeks. The full experience remains exclusive to the desktop app, with cloud-based sessions allowing for cross-device continuity.

Solos has unveiled the AirGo A6, a lighter version of its camera-less smart glasses, weighing around 19 grams and featuring an AI assistant that relies on voice interactions. The new model achieves weight savings through thinner temple arms and is available in multiple designs. Pricing and availability have not been finalized.

Solos has introduced two new pairs of smart glasses, including the camera-enabled AirGo V2, which can be paired with a separate Privacy Kit to block the camera for added privacy. The AirGo V2 offers features similar to Meta's smart glasses, including photo and video capture and an AI-powered assistant. The company's focus on privacy comes as Meta faces criticism for its own smart glasses, which have been called 'creepy' and have raised concerns about face recognition and data collection.

California dairy farmers are adopting a new approach to manure treatment using worms and microbes to reduce pollution, with several farms already implementing the vermifiltration system. This method aims to cut down methane, nitrous oxide, and water pollution generated by large amounts of manure. The state has established programs to promote the adoption of such methods, with over a billion dollars funneled to farms so far.