Meta rolls out Muse, a new AI image generator

Title: Tech Giants Expand AI Tools While Cutting Costs

01Meta Launches Muse Image Generator, Raises Data Consent Questions

Meta has rolled out Muse, a new AI image generation model targeting advertising, content creation, and creator tools. The launch represents Meta's most significant push yet into AI-generated imagery, with applications spanning from ad creative to home decoration. Muse is designed to integrate with Meta's existing advertising infrastructure, potentially giving businesses a new way to rapidly generate visual content at scale.

The launch has immediately sparked controversy over training data practices. According to Wired, Instagram users with public accounts will automatically have their photos included in Muse training data unless they actively opt out. The change means millions of users' images could be used to train the model without explicit consent, drawing criticism from privacy advocates and prompting questions about the future of AI training data practices across the industry.

02Claude Cowork Expands to Mobile and Web, Enabling Cross-Device Coding Agents

Anthropic has expanded Claude Cowork to mobile and web platforms, marking a significant escalation in the coding agent wars. The update allows developers to start a task from their desktop, receive status updates on their phone, and retrieve finished output later—even with their laptop closed. This cross-device capability represents a fundamental shift in how AI coding assistants can be used, moving from session-based interactions to persistent background agents.

The expansion signals a broader industry trend toward smartphone-controlled AI agents. Wired reports that Claude Cowork now keeps working on tasks even after users close their laptop, part of a larger push to make AI assistants truly persistent across devices. This positions Anthropic to compete directly with other coding agents like GitHub Copilot and Cursor, which have traditionally been tied to desktop development environments.

03Microsoft Cuts AI Spending by Relying on Own Models

Microsoft has joined the broader industry trend of cutting AI spending by depending more heavily on its own models. The company is the latest Silicon Valley giant to reduce external AI infrastructure costs by building and deploying proprietary models rather than relying on third-party services. This shift reflects mounting pressure on tech companies to demonstrate efficient AI spending as compute costs continue to climb.

The strategic pivot marks a significant change in enterprise AI procurement. By leveraging internal models, Microsoft can reduce per-query costs while maintaining tighter control over data and model behavior. The move aligns with similar strategies from other major tech companies, suggesting that the initial wave of heavy spending on external AI services may be giving way to a more cost-conscious approach emphasizing in-house development.

Also today

04
sqlite-utils 4.0 introduces database schema migrations — The latest major release of the Python utility library adds long-awaited database migration support alongside nested transactions and compound foreign key capabilities.
05
Hugging Face Models on Foundry Managed Compute — Microsoft's integration with Hugging Face enables seamless deployment of open-source models on Azure's managed compute infrastructure.
06
Chat Control passed first round in EU Parliament — The controversial EU legislation requiring scanning of private messages for illegal content advanced in parliamentary voting despite privacy concerns.
07
Discord admits AI moderation bug wrongfully banned users over harmless images — Discord confirmed that a bug in its AI moderation system has been incorrectly banning users since May, affecting hundreds of accounts.
08
OpenAI's Chief Futurist Is Leaving the Company — Joshua Achiam, who spent nearly nine years at OpenAI focusing on AI safety and notably testified in the Musk v. Altman trial, is departing the company.
09
Savi's app aims to protect consumers from realistic AI scams — Consumer protection startup Savi raised $7 million in seed funding to launch an app defending users against AI-generated voice scams like fake kidnapping demands.
10
Expanding Managed Agents in Gemini API — Google announced new capabilities for Managed Agents in the Gemini API, enabling developers to build production-ready AI agents with background tasks and remote MCP support.
11
Australian Payments Plus moves faster with ChatGPT and Codex — OpenAI highlighted how Australian Payments Plus uses ChatGPT Enterprise and Codex to accelerate payment processing while maintaining human oversight.