This website uses cookies

Read our Privacy policy and Terms of use for more information.

Recaply AI

Good morning, AI enthusiasts. OpenAI just launched Instant Checkout, letting ChatGPT users buy from over a million Shopify and Etsy merchants without leaving the chat.

Will AI assistants replace traditional search and marketplace platforms as the primary way we discover and purchase products?

In today's recap:

  • Shop and pay directly within ChatGPT conversations

  • Anthropic's new model codes continuously for over 30 hours

  • California enacts first US law mandating AI safety transparency reports

  • 8 new AI tools and more

OPENAI

ChatGPT now lets you shop without leaving

ChatGPT now lets you shop without leaving

Source: OpenAI

Recaply: OpenAI just made online shopping faster. ChatGPT users in the US can now buy products from Etsy and Shopify sellers without clicking away to another site. The new Instant Checkout feature turns the AI chatbot into a virtual store where shoppers can browse, compare, and purchase items all in one conversation.

Key details:

  • Instant Checkout works for US ChatGPT users across Free, Plus, and Pro plans with Etsy sellers now live and over one million Shopify merchants including Glossier and Skims rolling out soon.

  • Shoppers can complete purchases using Apple Pay, Google Pay, Stripe, or credit cards directly within their ChatGPT conversations without additional fees.

  • OpenAI built the feature with Stripe to handle secure transactions while merchants manage orders and fulfillment through their existing systems.

  • The company open-sourced its Agentic Commerce Protocol so other developers and merchants can integrate similar checkout experiences into their own platforms.

  • ChatGPT surfaces products based on relevance rather than paid partnerships, and OpenAI charges merchants a small transaction fee for completed sales.

Why it matters: This launch signals a shift in how people discover and buy products online. Instead of searching Google or browsing Amazon, shoppers can have natural conversations about what they need and complete purchases without switching apps. For OpenAI, transaction fees create a new revenue stream as the company works toward profitability. The move challenges traditional e-commerce giants by positioning AI chatbots as the new shopping assistants.

ANTHROPIC

Claude Sonnet 4.5 sets new coding bar

Claude Sonnet 4.5 sets new coding bar

Source: Anthropic

Recaply: Anthropic just released Claude Sonnet 4.5, and developers are excited. This new AI is better at coding, helps build AI agents, and can create software in real-time for subscribers. Anthropic keeps the same prices: $3 per million input tokens and $15 per million output tokens.

Key details:

  • Claude Sonnet 4.5 scores 77.2% on coding tests and can work for over 30 hours straight, much better than the old model's 7 hours.

  • The new Claude Agent SDK lets developers build custom AI helpers using the same tools that power Claude Code.

  • Max subscribers can try "Imagine with Claude," which creates software on the spot without pre-written code.

  • GitHub Copilot, Cursor, and Windsurf are already using Sonnet 4.5 in their tools.

  • The model is safer with better protection against tricks and deception.

Why it matters: Anthropic's plan is clear: give developers everything they need to experiment with AI agents. The big question is whether companies will use these tools enough to make them worth the cost.

AI SAFETY

First US AI safety law passes

First US AI safety law passes

Source: Getty Images

Recaply: California Governor Gavin Newsom signed the nation's first frontier AI transparency law on September 29. The Transparency in Frontier Artificial Intelligence Act (SB-53) requires major AI companies like OpenAI and Google to disclose their safety protocols and report incidents. Senator Scott Wiener authored the bill after his previous attempt was vetoed last year. This version takes a lighter approach, focusing on transparency rather than strict controls.

Key details:

  • The law targets companies with over $500 million in revenue building powerful AI models. These firms must publicly share safety frameworks and testing methods starting January 2026.

  • Companies report critical safety incidents to California's Office of Emergency Services within 15 days, or 24 hours if there's immediate danger to the public.

  • Whistleblower protections shield employees who raise safety concerns. The Attorney General can enforce civil penalties for companies that ignore the rules.

  • CalCompute launches as a public cloud platform giving startups and researchers affordable access to AI computing power for innovation.

Why it matters: California sets the standard as 32 of the world's top 50 AI companies operate there. This balanced approach aims to protect communities while keeping innovation alive. The law represents a middle ground between unrestricted development and heavy-handed regulation, potentially influencing federal and international AI policy.

NEWS

Beyond the big stories in AI today

  • Lovable launched Lovable Cloud and AI, enabling developers to build fullstack, AI-native apps with built-in backend, authentication, and file storage through simple prompts.

  • Cursor introduced two notable features, Cursor Browser and Cursor Learn.

  • DeepSeek introduced V3.2-Exp using sparse attention technology, cutting API costs in half for long-context operations through intelligent token selection.

  • Mintlify launched Agent for Slack, automating documentation updates from pull requests and code changes without requiring developers to open terminals.

  • Microsoft released Agent Mode in Excel, achieving 57.2% accuracy on SpreadsheetBench by combining advanced reasoning with Excel's calculation engine.

  • Rebellions raised $250M at a $1.4B valuation backed by Arm and Samsung, accelerating mass production of its energy-efficient REBEL-Quad AI inference chips.

  • OpenAI is preparing to launch a TikTok-like app for Sora 2, featuring vertical AI-generated videos limited to 10 seconds with facial verification for user likenesses.

  • Brave added Ask Brave to its AI search suite, delivering detailed report-style answers with contextual enrichments, generating over 15 million daily AI responses.

TOOLS

Trending AI Tools

  • 🧠 Fabric - Self-organizing workspace for your files and ideas

  • 🤖 Chatbase - Build AI chatbots trained on your website

  • 🎨 Leonardo AI - Generate stunning artwork and illustrations using simple text prompts

  • 📝 Glasp - Highlight, organize and share insights from articles and research papers

  • 🎤 Voicemod - Transform your voice in real-time with AI-powered voice effects

  • ✨ Adcreative - Generate high-converting ad creatives for social media marketing

  • 🧠 GitMind - Create mind maps and flowcharts fast with AI assistance

  • 🖼️ Pikzels - Create viral-quality YouTube thumbnails and titles in seconds with AI

🧡 Enjoyed this issue?

🤝 Recommend our newsletter or leave a feedback.

Cheers, Jason

Connect on LinkedIn, & Twitter.

Reply

Avatar

or to participate