Pinterest has launched an experimental AI shopping app called 'Ask Pinterest,' designed to help users discover products through natural language queries.

The app aims to provide a personalized shopping experience by understanding users' specific needs and preferences. It leverages Generative AI to interpret user queries and suggest highly relevant products, moving beyond traditional keyword searches.

'Ask Pinterest' is envisioned as an AI agent that not only assists with product discovery but also inspires users and streamlines their purchasing decisions. This initiative underscores Pinterest's ambition to strengthen its position as a practical shopping platform, complementing its visual discovery roots.

👉 Read the full article on TechCrunch

Reports indicate that the U.S. government's ban on Anthropic's AI models for foreign nationals was not primarily due to concerns over an 'AI jailbreak,' as widely speculated.

Instead, the prohibition likely stemmed from broader security concerns, issues related to access to sensitive data, or policy decisions impacting national security. This suggests that data privacy and the risk of information leakage in the use of LLMs were critical considerations for government agencies.

The move highlights the stringent standards governments are imposing on the safety and reliability of advanced Generative AI models as their use expands. It particularly suggests that access by foreign nationals is perceived as a potential risk.

👉 Read the full article on TechCrunch

Microsoft is reportedly considering using a self-hosted, modified version of China's open-source AI model, DeepSeek-V4, for its Copilot Cowork to reduce costs.

This consideration comes as models from Anthropic and OpenAI have become more expensive, pushing Microsoft to shift from an 'all-you-can-eat' to a usage-based pricing model. It aligns with Satya Nadella's vision of fostering a diverse LLM ecosystem, moving away from reliance on a few major players, as articulated in his essay, 'A frontier without an ecosystem is not stable.'

However, this move could provoke friction with the Trump administration, which has increased scrutiny on Chinese AI companies and considered banning foreign AI models like DeepSeek. While Microsoft aims for cost efficiency and AI ecosystem diversification, it faces potential political backlash.

👉 Read the full article on Gizmodo