This week in AI, commerce, creativity, and entertainment are evolving faster than ever. From AI-powered shopping inside ChatGPT to Anthropic’s most advanced coding model, and even the debut of the first AI actress, here’s what’s making headlines:

🛒 OpenAI Instant Checkout
ChatGPT now lets over 700M weekly users purchase products directly in chat via Etsy and soon Shopify, powered by the new open-source Agentic Commerce Protocol for secure, seamless payments.

⚡ Claude Sonnet 4.5
Anthropic’s upgraded model brings breakthroughs in coding, reasoning, and long-running agent workflows, offering developers new power with advanced safeguards and cost efficiency.

🎬 Tilly Norwood: AI Actress
Xicoia’s first AI-generated actress has already made her industry debut, attracting agency interest and sparking discussion about the role of synthetic talent in Hollywood’s future.

From smarter shopping to stronger coding and AI stepping onto the big screen, these updates highlight how fast AI is reshaping everyday experiences and industries alike.

ChatGPT Becomes Your Storefront: Shop Seamlessly in Chat

OpenAI has launched Instant Checkout for ChatGPT in the U.S., enabling over 700 million weekly users—including Plus, Pro, and Free tiers—to buy products directly in chat from Etsy sellers, with Shopify merchants joining soon; this purchase system uses the new open-source Agentic Commerce Protocol (built with Stripe), allowing fast, secure, and seamless single-item checkout while keeping merchants in control of payments and customer relationships—merchants pay a small fee per completed sale, but prices and product results remain unaffected and unsponsored, and Instant Checkout supports existing Stripe setups or other processors via flexible APIs, ensuring orders, payments, and fulfillment work with current merchant systems, minimal required data sharing, encrypted payment tokens, and explicit user confirmation, marking a move towards AI-driven, trustworthy commerce experiences that will soon expand to support multi-item carts, more merchants, regions, and developer integrations.

Claude Just Launched Upgraded AI Model: Sonnet 4.5

Claude Sonnet 4.5 is Anthropic’s most advanced AI model for coding, complex agent building, and computer use, showing major improvements in reasoning, math, and domain-specific expertise across finance, law, medicine, and STEM. It supports over 30 hours of continuous coding and task execution, featuring upgraded tools like checkpoints for saving and rolling back code, enhanced context editing, extended memory capabilities for longer agent runs, and integrated code execution with document creation in conversations. The model is highly aligned with strong safeguards against harmful outputs and prompt injection, available via Claude API, developer platforms like Amazon Bedrock, and integrated with tools such as GitHub Copilot. Claude Sonnet 4.5 offers cost-efficient pricing, state-of-the-art coding performance, and the new Claude Agent SDK to let developers build advanced AI agents with ease.

First AI Actor Becoming the Next Scarlett Johansson

Tilly Norwood, the first AI-generated actress created by AI talent studio Xicoia, made her debut at the Zurich Summit during the Zurich Film Festival’s industry segment. Created by actress, comedian, and technologist Eline Van der Velden, Tilly has attracted significant attention from talent agencies, indicating growing acceptance of AI performers in entertainment. Van der Velden highlighted how initial skepticism has shifted to collaboration interest, noting audiences focus on storytelling regardless of whether the actor is AI or human. Tilly’s first on-screen role was in a comedy sketch titled "AI Commissioner," which humorously explores the future of television. The AI talent studio expects to soon announce which agency will represent her, marking a notable step in the rise of synthetic actors in Hollywood.

Hand Picked Video

Create runway-ready fashion videos in seconds! With AI Fashion Video Background Blur, highlight your models and outfits while hiding distractions. Batch edit clips, blur faces, and export in HD for reels, campaigns, or e-commerce, completely free. Make every fashion shoot look studio-quality, anywhere!

Try Now - BGBlur 

Top AI Products from this week

• C1 by Thesys - C1 is a Generative UI API that augments LLMs to respond with interactive UI like charts, forms, cards, and more, in real time. It integrates seamlessly with any LLM, framework, or MCP in just 2 lines of code. Ship AI apps 10X faster with 80% less UI overhead.

• Assembly – Create remarkable customer experiences with Assembly (formerly Copilot). Get an out-of-the box branded client portal where customers can message you, pay bills, and share information securely, with a new AI-powered CRM for your team.

• Station – It's hard to monetize as a podcaster and YouTuber. Most make 70% less than other creators. Station uncovers hidden sponsor revenue and helps you activate all revenue — while you focus on episodes.

• Dex - Dex is the AI Headhunter that works for you. Tell Dex your needs and ambitions, and he'll scour 10,000s of jobs daily to connect you with the best roles at the world's leading AI Labs, Hedge Funds and Startups. Dex is the last recruiter you'll ever need.

• Alter for Meetings - Alter turns meetings into clear action—without a bot. It auto starts/stops recording for online meetings, transcribes and labels speakers locally for speed and privacy, then lets you chat using local models or cloud models for extra intelligence.

This week in AI

• ChatGPT Safety for Families - OpenAI lets parents link accounts with teens to set usage limits, quiet hours, disable voice, memory, image generation, and receive safety alerts for a safer AI chat experience.

• Amazon's AI Ad Partner - Amazon Ads' new agentic AI tool creates pro-quality ads fast via chat, using retail insights to design videos, images, scripts & voiceovers for brands

• Kling AI Studio - Kling AI is a next-gen creative studio generating custom brand images and videos via AI, supporting content teams with asset management and diverse style generation.

• DeepSeek V3.1-Terminus Update - DeepSeek V3.1-Terminus boosts agent tool reliability, reduces language glitches, and improves multi-step reasoning, offering more stable, consistent AI outputs.

• Apple Veritas AI - Apple’s internal chatbot Veritas tests new Siri AI features like personal data search and photo editing, aiming for a revamped, more reliable Siri by March 2026.

• Meta’s CWM Model - Meta’s 32B-parameter CWM LLM innovates code generation by modeling code execution, enabling neural debugging, agentic coding, and reasoning with a 131k token context window

Paper of The Day

IRG: Iterated Reasoning Guidance for Text-to-Image Generation" presents a novel approach to improving text-to-image synthesis by decomposing the complex generation task into explicit iterative reasoning steps. This method involves first generating a text-based plan to guide an initial image synthesis, followed by a reflection step that produces textual refinements which steer the final image generation. Trained on the IRGL-300K dataset, this two-stage framework significantly advances fidelity, fine-grained detail, and overall quality, achieving state-of-the-art results with a GenEval score of 0.85. The approach demonstrates that breaking down generation into a "reason-generate-reflect-refine" loop enhances controllability and outcome precision in generative models, marking an important progression for AI image synthesis techniques. This framework benefits AI practitioners by providing a scalable means to improve the quality and coherence of generated images through structured reasoning processes.

To read the whole paper 👉️ here.