OpenAI wants you starting mornings with Pulse

Hello Readers👀 

Curious about the biggest AI moves today? You’re in the right place. In today’s edition, we have:

🌅 OpenAI Unveils ChatGPT Pulse for Personalized Morning Briefs

🎨 From Figma to Code with Claude Code

🤖 GPT-5 Edges Closer to Human-Level Work

🔨Tools and Shifts you cannot miss

🌅 OpenAI Unveils ChatGPT Pulse for Personalized Morning Briefs

OpenAI has introduced ChatGPT Pulse, a new feature designed to deliver curated morning briefings while users sleep. Starting with Pro subscribers, Pulse is a proactive step toward making ChatGPT a true AI assistant, offering tailored updates, event agendas, and personalized content without needing a user prompt. 

The Shift:

1️⃣ Personalized AI Morning Reports - Pulse creates 5–10 dynamic cards overnight, combining data from past chats, user feedback, and optional integrations with Google Calendar or Gmail. Each card includes AI-generated images and concise summaries, covering anything from news about favorite sports teams to personalized travel itineraries.

2️⃣ Built for Efficiency and Balance - Reports conclude with a “Great, that’s it for today” message, a deliberate design choice to avoid engagement-driven scrolling. Users can request new automated reports or give quick feedback to refine future briefings, blending productivity with mindful consumption.

3️⃣ Future-Ready Agentic Capabilities - OpenAI plans to expand Pulse to Plus subscribers and evolve it into a more agentic assistant, capable of making restaurant reservations or drafting emails for approval. 

Pulse merges personalization, productivity, and proactive AI into one daily touchpoint, positioning ChatGPT as a morning go-to for news and planning. By linking to user apps and learning preferences, it highlights the future of AI as an always-on personal aide that saves time and streamlines the day before it even begins.

🎨 From Figma to Code with Claude Code

Design handoff doesn’t need to be messy. With Claude Code’s new Figma integration via MCP, you can turn mockups into clean, production-ready code in minutes, keeping design tokens, layouts, and hierarchy intact.

Here’s how to do it:

Step 1: Install the Figma Connector: Download the Claude desktop app and install the Figma Connector. This lets Claude “see” your design at the component level.

Step 2: Enable Dev Mode in Figma: Open the Figma desktop app → Preferences → toggle Enable Dev Mode MCP Server. This activates the link between your design and Claude.

Step 3: Connect Claude to Figma: Run this in your terminal:

claude mcp add --transport sse figma-dev-mode-mcp-server http://127.0.0.1:3845/sse
This bridges Claude and your Figma file.

Step 4: Share Your Design: Copy your Figma design link and paste it into Claude. Tell it which framework you want (React, Vue, Tailwind, etc.).

Step 5: Generate Code: Claude analyzes hierarchy, tokens, spacing, and interactivity → then outputs semantic, production-ready code.

💡 Why it matters: Instead of exporting static specs, you’re handing devs actual working code. This kills the back-and-forth and preserves brand consistency straight from design to production.

🤖 GPT-5 Edges Closer to Human-Level Work

OpenAI’s new GDPval benchmark is designed to test how frontier AI models perform on real, economically valuable work. By comparing AI-generated outputs to those of seasoned professionals across major industries, the results show just how rapidly systems like GPT-5 are catching up to human expertise.

The Shift:

1️⃣ Comprehensive Benchmark Across Industries - GDPval evaluates 1,320 tasks spanning 44 occupations in nine key sectors such as healthcare, finance, manufacturing, and government. Each task was crafted by professionals averaging 14 years of experience, providing a rigorous standard for AI to match or exceed human-level quality.

2️⃣ Standout Model Performance - Claude Opus 4.1 achieved the highest overall score with a 47.6% win rate, excelling in visual presentation tasks, while GPT-5 led in technical accuracy and matched or outperformed expert work in 40.6% of tasks. Other tested models included Gemini 2.5 and Grok 4, highlighting an increasingly competitive field.

3️⃣ Rapid Gains in Capability - OpenAI reports that GPT-5’s results are nearly three times higher than GPT-4o’s 13.7% recorded just 15 months ago, signaling a steep upward trajectory and significant progress in handling complex professional tasks.

GDPval shows that top AI models are no longer confined to narrow benchmarks: they’re moving toward practical, high-value applications across industries. While not replacing entire jobs yet, GPT-5 and its peers can already offload routine analysis, helping professionals focus on higher-impact work as AI capabilities accelerate.

🔨AI Tools for the Shift

🎶 FreeMusicGen  – Generate personalized music with lyrics or instrumentals powered by AI, instantly turning your ideas into full songs.

💊 Amino: Supplement Scanner – Scan, analyze, and track supplements for safety with this smart AI-powered scanner.

🕵️ Chrome Sidekick  – Unlock superpowers inside your browser with an AI agent that automates workflows, explains concepts, and extracts data.

🏥 Hathr AI – HIPAA-compliant AI tool powered by Claude that automates document summaries, creates analyses, and generates professional letters with government-grade security.

🤝 Nonverbia – Reads between the lines of every sales interaction, decoding nonverbal signals to uncover resistance and reveal deal-winning insights.

🚀Quick Shifts

🤖 Elon Musk’s xAI will sell Grok to U.S. agencies for just $0.42 per 18-month license, far cheaper than OpenAI and Anthropic’s $1 offers, aiming to win government AI contracts.

🎵  Spotify will label AI-generated tracks using the DDEX standard, ban unauthorized voice clones and deepfakes, and deploy a new spam filter to curb low-quality AI music uploads.

💰 Anthropic’s historic settlement could pay authors about $3,000 per book or work used to train its AI models, marking the first large-scale compensation for creatives whose content powered AI systems.

🤖 Google DeepMind’s new Gemini Robotics 1.5 and Gemini Robotics-ER 1.5 let robots plan multiple steps ahead, separate laundry, pack weather-ready suitcases, and even consult the web to sort local recycling.

🎥 Meta introduced Vibes, a TikTok-style feed inside the Meta AI app and meta.ai that features only AI-generated short videos. Users can create, remix, add music, and cross-post to Instagram or Facebook.

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇