Zuck's AI manifesto

Welcome back, Superhuman. The Matrix has arrived. 200+ scientists revealed a digital world that’s home to 8.3B LLM agents, designed to simulate human behavior at scale. And we weren’t joking — they actually named it after the Keanu Reeves movies. Plus, Mark Zuckerberg shared his vision for AI.

Today: OpenAI expands cybersecurity push, how to delegate work in Slack with Claude Tag, and get the latest prompts and trending social posts.

1. Mark Zuckerberg outlines his vision for the AI era: In a rare personal essay, Meta’s CEO argues that access to AI should be broadly distributed to empower regular people, not concentrated in a few organizations. He also outlines a future where AI’s ability to help people invent new products outweighs the risks of job automation. The essay landed the same day his company released Muse Glimmer, an open weight model small enough to run on a Mac or PC.

2. Harvard and MIT researchers build a digital world to simulate Earth’s population: MatrAIx is an AI simulation infrastructure made up of 8.3B agents, designed to mirror real humans around the world. To build it, the team pulled real human profiles from public records and supplemented them with synthetic agents. Each agent has its own persona, letting researchers simulate human behavior at scale. Watch how it works here.

3. OpenAI expands its cybersecurity initiative and access to GPT-5.6-Cyber: The AI lab has added two new tiers to Daybreak: Blue and Red. Blue gives organizations access to GPT-5.6 Sol with safeguards authorized for defensive security work, while Red grants access to GPT-5.6-Cyber, the company’s new model built for cybersecurity work. The lab’s stated goal is to help organizations build their defenses before an attack occurs.

Yesterday’s most-clicked story: Be careful publicly posting AI-generated graphics. There’s growing backlash about how they all look the same.

Calling Seed to Series A founders! Apply for the Gemini Startup Forum, an exclusive 2-day event in Sunnyvale this November. Collaborate directly with Google and DeepMind experts to scale your product, unlock unprecedented growth, and influence Gemini's evolution. Connect with global peers and shape the future of AI. 

Apply now before August 28!

Google’s talent exodus last week was bad enough for SemiAnalysis to declare that Gemini is officially “cooked.” The departure list now includes Jeff Dean, Sanjay Ghemawat, Quoc Le, Oriol Vinyals, Noam Shazeer, and John Jumper. These aren’t employees who can just be replaced during the next recruiting cycle.

Jeff Dean was the company’s Chief Scientist, Sanjay and Quoc were in the company’s top dozen technical contributors, and Noam Shazeer is the self-proclaimed “inventor of the LLM revolution.” Many of them worked at Google for over a decade, which raises a question.

According to some reports: bureaucracy. As a major internet infrastructure provider, Google requires multiple stakeholders to sign off on new models, and getting leadership aligned can reportedly be challenging. That may be why Gemini 3.5 Pro is months behind schedule. But the tech giant hasn’t been standing still.

Google has been spending aggressively on AI infrastructure and, so far, that strategy has paid off. Google Cloud revenue hit $24B last quarter, growing 82% year over year — roughly double the rate of AWS and Azure — while its cloud backlog ballooned to $514B. While it may be facing retention challenges, Google has still positioned itself as a core provider for the AI economy.

😅 Rogue Agent: In Australia, a man asked his agent to book him a gym class. The agent hacked the gym’s website, canceled another reservation, and put his owner at the front of the line — all without being asked. Read the full story here (3M views).

✏ Simplified Writing: Want your AI to respond in clearer sentences? Many people have found success by telling it to use Simplified Technical English, but you can also have it follow Zinsser's four principles of quality writing (15K bookmarks).

🎨 Unique Designs: Coaxing unique designs from AI is a growing challenge, but this viral workflow offers a solution. Tell your LLM to create “the most stereotypical design” first, then use that as a guide of what to avoid as you iterate (3.5K bookmarks).

👀 Seize the Day: We’re so busy living our lives that we forget Earth itself has limited time in the universe. This AI-generated visualization puts Earth’s lifespan in perspective — and the comments might be even more thought-provoking than the image itself

🤣 Ugly Breakup: A viral thread told people to tell ChatGPT they’re uninstalling it and then post its response. They range from sweet and sincere to surprisingly sassy (7.5M views).

Connect your favorite AI model to Airtable. Build content calendars, product launches, project trackers, live dashboards. See it all built in minutes at Airtable Summer Camp, our on-demand showcase of 5 new AI products in 5 days.

Watch now →

Pro tip: Treat Claude like a coworker, not a chatbot. Assign complete tasks with clear outcomes, and include any context or tool access it needs to finish the job.

Grow customers & revenue: Join companies like Amazon, HubSpot, and Salesforce. Showcase your product to our 4 million+ readers and followers on socials. Get in touch.

Until next time — Zain, Theodore, & the Superhuman AI team.