• The Current ⚡️
  • Posts
  • Meta Releases Standalone AI Chatbot to Directly Compete With OpenAI

Meta Releases Standalone AI Chatbot to Directly Compete With OpenAI

Also, MasterCard reveals “Agent Pay” for AI powered payments

Author

Jack Lajoie
April 30, 2025 • read time ~ 9 minutes

⚡️ Headlines

🤖 AI

Trump-aligned policymakers may overhaul Biden’s AI chip export controls, favoring bilateral licenses over current restrictions. [Reuters]

Lyft introduces an AI-powered assistant designed to help drivers optimize their time, routes, and income. [The Verge]

Google’s NotebookLM adds support for audio summaries across 50+ languages, enhancing accessibility for global users. [Google Blog]

A growing tension between OpenAI CEO Sam Altman and Microsoft’s Satya Nadella centers on product vision and strategy alignment. [WSJ]

Kintsugi, which automates sales tax processing using AI, sees its valuation double amid strong enterprise demand. [TechCrunch]

Reddit has shut down an AI project that replicated user behaviors without permission, citing privacy violations. [The Verge]

Meta restricts underage users from accessing its AI Studio on Instagram, citing safety concerns. [404 Media]

Agent Pay by Mastercard brings AI-powered search to enterprise workflows, eliminating context-switching for employees. [VentureBeat]

Nvidia’s new AI tool generates 3D scenes from text prompts and Blender image references, targeting creators and developers. [The Verge]

🦾 Emerging Tech

Hugging Face introduces an affordable 3D-printed robotic arm aimed at developers and hobbyists. [TechCrunch]

Meta’s updated privacy policy reveals its smart glasses now collect voice data for training AI models. [The Verge]

Bitcoin edges near $95,000 with equities holding firm, though analysts warn of speculative momentum. [CoinDesk]

🤳 Social Media

YouTube is experimenting with blurring thumbnails for flagged adult content to reduce accidental views. [Social Media Today]

WhatsApp announces AI tools that analyze messages without compromising user privacy through on-device processing. [Meta Engineering]

🔬 Research

A new report from Common Sense Media rates AI social companions on safety, transparency, and developmental impact. [Common Sense Media]

UC San Diego researchers use AI to uncover cellular processes linked to Alzheimer’s, pointing to a potential treatment path. [UCSD Today]

🎱 Random

A hands-on review of Rabbit’s AI wearable explores its real-time task execution and natural-language interface. [WSJ]

Amazon refutes a report alleging it inflated prices in response to tariff shifts, calling it inaccurate. [Fortune]

🔌 Plug-Into-This

Meta has launched a standalone AI assistant app powered by its Llama 4 model, aiming to rival ChatGPT. Distinctively, the app introduces a social component through its "Discover" feed, allowing users to view and interact with AI-generated content shared by friends on Instagram and Facebook. This integration seeks to demystify AI by showcasing its applications within a familiar social context.

  • Social Integration: The "Discover" feed enables users to like, comment on, share, or remix AI interactions shared by their social network, fostering a community-driven exploration of AI capabilities.

  • Voice Interaction: An experimental "full-duplex" voice mode offers dynamic, conversational engagement, enhancing the naturalness of AI interactions.

  • Personalization: By linking Facebook and Instagram accounts, the app personalizes responses, tailoring interactions based on user data from these platforms.

  • Multimodal Capabilities: Beyond text and voice, the assistant supports image generation and real-time web search, broadening its utility.

  • Hardware Synergy: The app replaces the former View app for Meta’s Ray-Ban smart glasses, integrating AI features like object recognition and real-time translation.

🧠 Meta's approach intertwines AI functionality with social media dynamics, positioning its assistant not just as a tool but as a shared experience. This strategy leverages existing user behaviors to accelerate AI adoption, potentially setting a precedent for future AI-social integrations.

Mastercard has introduced Agent Pay, an innovative payment solution designed to integrate seamlessly with AI-driven commerce. This initiative focuses on enabling autonomous agents to execute transactions on behalf of users, emphasizing trust, security, and control in the evolving landscape of AI-assisted commerce.

  • Agentic Tokens: Building upon existing tokenization technologies, Agentic Tokens facilitate secure, programmable payments, supporting scenarios like recurring expenses and subscriptions.

  • Collaborations: Partnerships with Microsoft and IBM aim to scale agentic commerce, leveraging technologies like watsonx Orchestrate to enhance B2B use cases.

  • Ecosystem Integration: Collaborations with acquirers and checkout platforms such as Braintree and Checkout.com seek to extend tokenization capabilities across merchant networks.

  • Banking Integration: For financial institutions, tokenized payment credentials will be integrated across agentic platforms, maintaining issuer relevance in AI-driven transactions.

  • User Experience: Agent Pay aspires to embed trusted payment experiences within AI-generated recommendations, streamlining the path from discovery to purchase.

💳 Mastercard's Agent Pay represents a strategic move to redefine payment infrastructures in anticipation of AI's growing role in commerce. By proactively developing systems that accommodate autonomous transactions, Mastercard positions itself at the forefront of agentic commerce innovation.

Meta has launched the AI Defenders Program, introducing a suite of open-source tools aimed at enhancing the safety and security of AI systems. This initiative underscores Meta's commitment to responsible AI development, providing resources to the broader community to address emerging challenges in AI deployment.

  • Llama Guard 4: An updated moderation model designed to filter harmful or inappropriate content, ensuring safer AI interactions.

  • LlamaFirewall: A security-focused guardrail framework serving as a final defense layer against potential threats in AI applications.

  • Llama Prompt Guard: Tools aimed at preventing prompt-based attacks, safeguarding the integrity of AI responses.

  • Community Engagement: By open-sourcing these tools, Meta encourages collaborative efforts to refine and enhance AI safety measures.

  • Proactive Safety Measures: The program reflects a proactive stance in addressing AI risks, emphasizing the importance of integrating safety at every development stage.

🛡️ Meta's AI Defenders Program highlights the growing recognition of AI safety as a collective responsibility. By equipping developers with robust tools, Meta fosters a collaborative environment where safety is integral to innovation, setting a standard for industry-wide practices.

 🆕 Updates

📽️ Daily Demo

🗣️ Discourse