🔌The USB-C for AI

Hey folks,

It is a Happy Friday when the AI industry's biggest rivals just agreed to play nicely together. OpenAI, Anthropic, Google, Microsoft, and AWS launched the Agentic AI Foundation, a coalition to standardize how AI agents connect to tools and data. The pitch: a "USB-C for AI" that prevents vendor lock-in and lets agents work across platforms. Developers can now connect agents to Google services by pasting a single URL instead of spending weeks building connectors.

Mistral also shipped a Vibe CLI, an open-source command-line coding assistant that scans your project, understands Git history, and orchestrates multi-file changes.

Let's dive in..

🔌 OpenAI and Anthropic Agree on How AI Agents Should Work

🐝 The Buzz: What happens when the two fiercest AI rivals decide to collaborate? The Agentic AI Foundation launched under the Linux Foundation with OpenAI, Anthropic, Google, Microsoft, AWS, Block, Bloomberg, and Cloudflare as founding members. Anthropic donated MCP, now the de facto standard for connecting AI agents to tools, while OpenAI contributed AGENTS.md, which 60,000+ repositories already use to tell coding agents how to behave.

The contributions are substantial:

• Anthropic's MCP has 97M+ monthly SDK downloads across TypeScript

• OpenAI's AGENTS.md has been adopted by Cursor, Devin, GitHub Copilot, Gemini CLI, and VS Code

• Block donated Goose, an open source agent framework used by thousands of engineers weekly

• Google joined as a platinum member alongside AWS and Microsoft

But here's what's different: This isn't a marketing alliance, it's infrastructure. MCP now has an official registry for discovering servers, asynchronous operations, and official SDKs in every major language. Anthropic's new "code execution with MCP" pattern reduced token usage by 98.7% in one workflow, from 150,000 tokens to 2,000. That's the kind of efficiency gain that makes production-scale agents economically viable.

Google made it real the next day: Managed MCP servers went live for Maps, BigQuery, Compute Engine, and Kubernetes Engine. Instead of weeks building custom connectors, developers paste a URL to a managed endpoint. Google Cloud tested it with Claude and ChatGPT as clients and they just work.

The skeptic's corner: Apple and Meta are notably absent from the coalition. Whether this becomes a true open standard or just the "Google/Microsoft/OpenAI" standard remains to be seen. But with 97 million monthly SDK downloads, MCP has critical mass.

💡 Takeaway: The agent infrastructure layer just consolidated. If you're building agents or evaluating platforms, MCP compatibility is now the norm. The days of building custom integrations for every service are ending and the companies that bet on proprietary agent protocols are about to look very lonely.

Together with Levanta

The Future of Shopping? AI + Actual Humans.

AI has changed how consumers shop by speeding up research. But one thing hasn’t changed: shoppers still trust people more than AI.

Levanta’s new Affiliate 3.0 Consumer Report reveals a major shift in how shoppers blend AI tools with human influence. Consumers use AI to explore options, but when it comes time to buy, they still turn to creators, communities, and real experiences to validate their decisions.

The data shows:

• Only 10% of shoppers buy through AI-recommended links

• 87% discover products through creators, blogs, or communities they trust

• Human sources like reviews and creators rank higher in trust than AI recommendations

The most effective brands are combining AI discovery with authentic human influence to drive measurable conversions.

Affiliate marketing isn’t being replaced by AI, it’s being amplified by it.

Download the full report to see what this means for your brand.

🎨 Adobe Just Brought Photoshop to 800 Million ChatGPT Users

🐝 The Buzz: Until now, you couldn’t edit a photo in ChatGPT. Adobe shipped Photoshop, Express, and Acrobat integrations to ChatGPT, available free to all 800 million users starting now. Type "Adobe Photoshop, blur the background of this image" and ChatGPT surfaces Photoshop's editing engine.

What you can actually do:

• Photoshop: Remove/blur backgrounds, adjust exposure and contrast, apply effects with intensity controls

• Express: Fetch designs from your library, create themed content, animate elements, edit existing work

• Acrobat: Merge PDFs, extract text and tables, edit documents inline

The reach is massive: ChatGPT desktop, web, and iOS get all three integrations immediately. Adobe requires account login for generation, turning ChatGPT into a massive lead-generation funnel for Creative Cloud.

Here's the competitive context: Canva and Figma landed in ChatGPT back in October when OpenAI launched its Apps SDK. Adobe's been playing catch-up, but "Photoshop" is still the verb people use for image editing. The integration builds on MCP, the same protocol that just got donated to the Linux Foundation. Adobe's betting conversational AI becomes the new creative interface.

💡 Takeaway: The creative workflow is fragmenting. Your team might edit images in ChatGPT, design in Figma, and export from Canva, all without opening a traditional app. Adobe's move signals that even the incumbents see AI assistants as the future interface for creative tools.

📰 Like newsletters? Here are more our readers enjoy.

💻️ How to Use AI For Sports Coaching

With Gemini 3’s massive long-context window and vision and spatial understanding, you can upload a video of yourself playing a sport for up to an hour and receive coach-level advice. Gemini 3 will identify that you’re the player, filter out noise and offer a detailed visual analysis, complete with information like form evaluation and suggested drills.

• Go to Gemini

• Upload your recorded game footage

• Prompt the following

Analyze this video. Act as a certified Pickleball coach. Watch this doubles match and provide a technical performance audit for the woman in the yellow hat and knee brace, identifying any areas of improvement and specific exercises or drills to correct them.

🐝 AI Buzz Bits

🧠 A Tokyo startup claims to have built the world's first AGI model. Integral AI, founded by ex-Googler’s say their system can learn new tasks without datasets or human intervention, one of three criteria they've defined for AGI alongside safe mastery and human-equivalent energy efficiency. The company demonstrated robots learning skills without supervision, claiming the architecture mirrors the brain's neocortex. No public benchmarks yet, and "world first" AGI claims deserve healthy skepticism, but this is the boldest declaration we've seen from a funded startup.

🥽 Google shipped three major Galaxy XR updates: PC Connect lets you pull Windows desktop windows into your headset workspace, Travel Mode stabilizes your view during flights, and Likeness creates a real-time digital avatar of your face for video calls. Google also previewed Project Aura from XREAL, the first Android XR glasses with a 70-degree field of view and optical see-through tech.

🔥 AI productivity tools

For a full list of 1000+ AI tools, visit our Supertool Directory

🤘 TGIF