ChatGPT learns to use Chrome

Welcome back, Superhuman. AI assistants are finally playing nice with Google’s most popular tools. Two weeks ago, Claude set up shop in Google Workspace. Now ChatGPT is moving into Chrome, where it can offer advice and take action anywhere across the browser.

Today: Anthropic launches a public fund, how to turn rough notes into a polished deck with Claude Code, and get the latest prompts and trending social posts.

1. OpenAI brings computer use to Google Chrome: A new update lets ChatGPT see what you’re looking at in Google Chrome to give you more relevant answers, with ChatGPT Work going one step further by taking action across your Chrome tabs (see examples of how you could use it). OpenAI also shipped two upgrades for Pro users: faster responses when using GPT-6 and multi-agent workflows, with individual tracking for each subagent. Try ChatGPT’s Chrome extension here.

2. Anthropic announces a public fund as agents drive productivity: In a new report, Chief Economist Peter McCrory says employee-directed agents worked 6x as many hours as their human users in August, driving productivity across the company. In response, the lab has slowed hiring for junior technical roles. Separately, Anthropic committed 5% of its shares to a public benefit fund designed to channel growth into public giving, pending shareholder approval.

3. Neocloud stocks slide amid rising borrowing costs: Shares of companies that provide AI cloud infrastructure — like CoreWeave, Nebius, and IREN — are in an intensifying sector-wide decline, with all three falling over 20% last week. Goldman Sachs suggests the selloff is due to the rising cost of funding these projects, which has led to increased investor skepticism about the economics behind the AI buildout.

Adding AI to your workflow is easy, but knowing if it’s actually increasing output? Much harder. 

Watch Atlassian’s State of AI SDLC summit (on demand) to learn from CEOs and leaders at Atlassian, Vercel, Dropbox, and more as they unpack how they assess AI ROI: 

Watch the full summit here.

Sample Prompt: “Turn the files in this folder into a polished [10]-slide deck for [audience]. Build a clear narrative from [problem] to [solution/outcome], keep each slide focused on one idea, and use concise copy. Add charts, diagrams, device mockups, and visual elements where useful. Create the final output as a Claude Slides artifact in a [minimal / premium / bold / corporate] style.”

😅 Top labs aren’t letting up. In the past month or so, we’ve gotten new models from Google, Anthropic, and OpenAI, not to mention Meta’s developer-first Muse family and each lab’s flagship variants. Here’s where each model is most effective.

Google’s most powerful model hit digital shelves last Tuesday, after a limited release in late September. CEO Sundar Pichai touted the model’s ability to use multiple tools and generate 3D assets, like this photorealistic pagoda temple that’s been circulating online. A DeepMind developer also highlighted its deep understanding of the natural world.

Many Anthropic fans were underwhelmed by Opus 5. Opus 5.5 is winning them back. The new model excels at long-running tasks, offers a straightforward writing style, and can create animations and motion graphics from code (no visual assets required). For your most ambitious projects, Anthropic's Fable 5.1 is another strong option.

Former Apple engineer Anshu Chimala called Astra a “machine god” for building 3D games, while others have leaned on it for interactive visualizations. It also handles computer tasks with ease. While everyone’s got their preferred option, it’s critical to keep experimenting — you never know when a new model will surprise you.

Your team of people and agents need the controls of business software at the scale of infrastructure. Attio gives them both:

Explore the new architecture of software here.

📈 Market Paradox: The Kobeissi Letter says investors are focusing too much on negative economic headwinds and overlooking AI-driven earnings growth. They argue this gap has created a “historic opportunity” for stock investors (4M views).

📺 Debloat Your TV: Hobbyists are using Claude Code to remove ads from their smart TVs, with dozens of viral posts talking about how fast their TV works after Claude strips it down. This step-by-step guide can help you do it (2.8M views).

🪰 Simulation Theory: The simulation hypothesis says our universe is one giant computer program. For one simulated fruit fly, that’s now its reality thanks to the world model lab Odyssey (4.7M views).

🚨 Watch Out: Hackers are sending prompt injections that you can’t see, but your email-summarizing agent can. Here’s what to watch out for.

🪄 Malfoy Mania: AI’s public sentiment may be starting to shift, but not because of powerful models or PR campaigns. It’s Harry Potter fan fiction starring Draco Malfoy as a girl that’s turning the tide, with the AI-generated clips growing more popular each week.

👉 Click here to open the full prompt in ChatGPT or Claude

👉 Click here to open the full prompt in ChatGPT (Don’t forget to switch to Images!)

Grow customers & revenue: Join companies like Amazon, HubSpot, and Salesforce. Showcase your product to our 2 million+ readers and followers on socials. Get in touch.

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