OpenAI’s e-commerce takeover

OpenAI Revolutionizes E-Commerce with Instant Checkout

OpenAI has launched Instant Checkout, a feature enabling U.S. users to complete purchases directly within ChatGPT conversations. This eliminates the need to navigate away from the chat interface, streamlining the shopping experience.

• Built in partnership with Stripe, Instant Checkout currently supports Etsy sellers, with plans to onboard over one million Shopify merchants soon.
• Users can select a product recommended by ChatGPT, review order details, and finalize payment-all without leaving the chat window.
• The underlying Agentic Commerce Protocol (ACP) has been open-sourced, allowing any retailer to integrate this seamless checkout with minimal coding, especially Stripe merchants.
• While OpenAI earns fees from completed sales, product rankings remain organic and relevance-based, ensuring unbiased recommendations.

This development marks a pivotal moment in AI-driven commerce, potentially opening a lucrative revenue stream for OpenAI as shopping increasingly shifts into conversational AI platforms. The future may also see the integration of advertising within this ecosystem, further evolving the digital retail landscape.

Partner Spotlight: Turing’s Collaborative AI Research Acceleration

In the race to advance AI, quality and collaboration trump sheer data volume. Turing partners with leading AI labs to co-own research outcomes and engineer sophisticated human-AI feedback loops, accelerating breakthroughs.

• Focuses on shared experimental results rather than just data delivery, maintaining vendor neutrality.
• Implements quality-centric workflows with transparent data lineage and auditable outputs.
• Develops custom reinforcement learning environments and pipelines tailored to specific benchmarks.

Turing’s approach addresses the nuanced needs of frontier AI research, bridging gaps between potential and state-of-the-art performance.

Anthropic Launches Claude Sonnet 4.5: A New Benchmark in AI Coding

Anthropic has introduced Claude Sonnet 4.5, hailed as the world’s leading coding AI model. It delivers state-of-the-art results on real-world software development benchmarks while maintaining the same API pricing as its predecessor.

• Achieves nearly 20% performance improvement over Opus 4.1 on computer usage tasks.
• Demonstrated autonomous coding for over 30 hours, generating 11,000 lines of code-significantly surpassing GPT-5-Codex’s 7+ hour coding sessions.
• New features include Claude Code checkpoints, enhanced memory and context editing via API, and a Claude Agent SDK for building autonomous agents.
• Launched “Imagine with Claude,” a 5-day research preview for Max users showcasing real-time software generation capabilities.

This release reasserts Anthropic’s position in the competitive AI coding space, hinting at a future where AI can handle complex, long-duration programming tasks with minimal human intervention.

How to Create Engaging Talking Head Videos Using AI

Want to produce professional avatar videos without filming? This step-by-step guide shows you how to generate a realistic talking head using Google Gemini and animate it with your voice via Wan Video.

1. Upload a photo to Google Gemini with the prompt: “Create a professional headshot of this person as a talking head, facing forward, wearing [desired outfit], with [background]. Close-up, professional lighting, high resolution.”
2. Open Wan Video, start a new project, and switch the media type from Video to Avatar.
3. Upload your Gemini-generated headshot and add audio by recording 10-15 seconds or typing up to 300 words using Wan’s built-in voices.
4. Click Generate to synchronize lip movements with your audio, then use the timeline to assemble multiple segments for a complete presentation.

Pro Tip: Prepare your script in advance and break it into short, natural pauses to ensure smooth, lifelike delivery.

Invisible’s Insight: Rethinking AI Evaluations for Real Impact

The AI community remains divided on the value of evaluations-some prioritize intuition, while others demand rigorous metrics. Invisible’s latest brief cuts through the debate, explaining why “benchmaxxing” can mislead and proposing smarter evaluation strategies.

• Clarifies the critical role of evaluations in bridging AI pilots and measurable ROI.
• Offers a practical framework for designing custom evaluations tailored to specific use cases.
• Details methods for constructing inputs, identifying flawed training data, and conducting behavioral and safety assessments.
• Highlights a client case where harmful behaviors were reduced by 97% using just 4,000 data rows instead of 100,000.

OpenAI’s Sora 2: A New AI-Driven Social Video Experience

OpenAI is developing Sora 2, a standalone social media app powered by its upcoming video model. Inspired by TikTok’s vertical feed, Sora 2 will exclusively showcase AI-generated short videos, capped at 10 seconds each.

• Includes identity verification to allow users to authorize the use of their likeness in generated videos.
• Copyrighted content will be permitted unless rights holders opt out; public figures must provide consent.
• Features remix tools and algorithmic recommendations similar to TikTok’s “For You” page, with notifications when a user’s image is used.
• Launch follows Meta’s introduction of Vibes, an AI-only video feed within its AI app, highlighting a growing trend toward AI-centric social platforms.

While promising, Sora 2 faces challenges to match competitors’ sophistication. Both OpenAI and Meta’s ventures into AI-driven social feeds have sparked mixed reactions, raising questions about the future quality and user experience of such platforms.

Quick Updates from the AI World

• Protect your online privacy by removing personal data from the web-use code RUNDOWN for 55% off*
• Anthropic’s Claude Sonnet 4.5 sets a new standard in AI coding performance
• Shop effortlessly during conversations with ChatGPT’s Instant Checkout
• Build AI apps faster with integrated backend services and Gemini-powered features

*Sponsored Listing

Additional Industry Highlights

• DeepSeek V3.2-Exp introduces a “sparse attention” mechanism, slashing API costs by over 50% while maintaining performance.
• California Governor Gavin Newsom signs SB 53, mandating transparency from AI companies, establishing a computing cluster consortium, and protecting whistleblowers.
• OpenAI debuts a safety routing system that switches to GPT-5-level reasoning during sensitive chats, alongside new parental control features.
• Quantum computing expert Scott Aaronson credits GPT-5-Thinking for a key technical breakthrough in his latest research paper.
• Lovable Cloud and AI platform enables users to create full-stack applications through prompts, integrating backend services with Gemini-powered AI.

Community Spotlight: AI in Healthcare

Each edition, we highlight how readers leverage AI to enhance their work and lives. Today, meet Tauseef M. from Bangalore, India:

“As a clinician, I rely on ChatGPT and Perplexity to swiftly analyze clinical data, stay updated with the latest guidelines, and craft personalized treatment and lifestyle plans. This boosts my efficiency and allows me to dedicate more time to meaningful patient care.”

How are you using AI? Share your story with us.

Explore More

• Read our previous AI newsletter
• Catch up on the latest Tech newsletter
• Discover insights in our Robotics newsletter
• Browse today’s AI tool guide
• RSVP for our next workshop, Friday at 4 PM