Good Morning! In a plot twist that sounds like 2024 tech-fiction, Anthropic just signed a massive compute deal with Elon Musk’s SpaceX to fuel its explosive growth.OpenAI Drops GPT-5.5 Instant. Plus, I’ll show you how to convert Gemini Nano Banana images to SVG using Google Slides.
Anthropic Taps SpaceX to Power the Claude Revolution
Zuckerberg Bets $500M to Build a "Virtual Cell"
Anthropic Gives Managed Agents "Dreams" to Self-Learn
How to Convert Gemini Nano Banana Images to SVG Using Google Slides
4 new AI tools worth trying
AI MODELS
Anthropic has officially secured exclusive access to Colossus 1, the massive Memphis data center built by xAI. Despite Elon Musk’s history of calling the company "woke" and "evil," the sheer scarcity of compute has forced the two rivals into a strategic marriage of convenience.
Anthropic gains immediate access to over 220,000 Nvidia GPUs (H100, H200, and GB200) and 300 megawatts of capacity, providing a "compute bridge" until its larger contracts with Amazon and Google go live in 2027.
The impact was instant: Anthropic has already doubled the five-hour rate limits for Claude Code (Pro, Max, and Enterprise) and removed peak-hour reductions for power users.
Anthropic’s growth is shattering internal records, moving from a projected 10x revenue increase to a staggering 80x, reportedly pushing its annualized revenue past OpenAI.
SpaceX is using this deal to pivot into a major AI infrastructure provider ahead of its planned IPO, with future plans to develop "orbital AI compute" via a network of data center satellites.
In 2026, compute scarcity has become the ultimate equalizer. When the largest AI lab in the world (by ARR) hits a hardware wall, even the deepest personal and political rivalries are set aside for GPUs. This partnership proves that the AI arms race is no longer about who has the best model, it's about who has the fastest access to power and silicon, whether it's on the ground in Memphis or eventually in low Earth orbit.
AI NEWS
Mark Zuckerberg and his wife Priscilla Chan’s nonprofit, the Chan Zuckerberg Biohub, just launched a five-year, $500 million Virtual Biology Initiative. The goal? To build predictive AI models of human cells to accelerate the cure and prevention of all diseases.
The initiative will use AI to create digital models of human cells, allowing scientists to run thousands of experiments virtually and predict how diseases develop before stepping into a lab.
$400 million is earmarked for Biohub’s internal AI development and advanced imaging tech, while the remaining $100 million will fund external researchers to help gather global biological datasets.
Biohub's Head of Science, Alex Rives, noted that accurately representing biological complexity requires "orders of magnitude more data than exists today," making this an enormous data-scaling challenge.
The massive philanthropic push arrives right as a new report from the Institute on Taxation and Economic Policy highlights that Meta paid a record-low federal income tax rate of just 3.6% on $79 billion in U.S. profits in 2025.
We are watching the exact same AI scaling laws that cracked language and coding be pointed directly at living biology. If Biohub can successfully gather enough data to create a functioning, predictive digital twin of a human cell, it would shift medical research from a slow, physical process to a lightning-fast digital simulation. It’s a moonshot engineering problem that, if solved, could fundamentally change the timeline for curing human diseases.
AI NEWS
Anthropic's newly launched Managed Agents platform just added a massive new feature: Dreams. Currently rolling out as a research preview via waitlist on the Claude Console, the feature serves as a self-learning solution that allows AI agents to evaluate past results and autonomously improve their future performance.
Sits directly within the "Manage agent context" stack alongside persistent memory stores, giving agents true long-term learning capabilities.
Allows agents to define outcomes and success criteria, self-evaluating past session runs and iterating on their own logic to meet those criteria more efficiently next time.
Operates securely within the Managed Agents harness, maintaining Anthropic’s strict enterprise-grade sandboxing, logging, and access controls.
Currently restricted to a research preview under a waitlist via the Claude Console, as developers begin testing its impact on complex, multi-step autonomous workflows.
We are officially shifting from agents that just "do tasks" to agents that actively learn from their mistakes. By giving Managed Agents the ability to "dream", reviewing past actions, recognizing failures, and self-optimizing for the next run, Anthropic is bridging the gap between static automation and truly intelligent, self-improving digital employees. If this scales, the AI agent you deploy on Friday will literally be smarter and more efficient by Monday morning.
HOW TO AI
🗂️ How to Convert Gemini Nano Banana Images to SVG Using Google Slides
In this tutorial, you will learn a surprising "first-party" hack to transform your raster Gemini Nano Banana images into high-quality Scalable Vector Graphics (SVG).
🧰 Who is This For
People turning AI images into editable vector graphics
Designers creating logos, icons, and illustrations
Content creators making clean scalable visuals
Developers needing SVG assets for websites/apps
STEP 1: Access the New "Gems Made by Labs"
First, download your Nano Banana image to your computer. To avoid major distortion, you need to match your workspace to the image.
Right-click your image file, select Properties > Details, and note the dimensions (e.g., 1792 x 2400). Now, open slides.google.com, start a blank presentation, and delete all default text boxes. Navigate to File > Page setup, select Custom, switch the units to Pixels, and enter your image's exact dimensions.
STEP 2: Import and Lock the Background
Navigate to Insert > Image > Upload from computer and select your AI-generated file. Once it appears on the canvas, right-click the image and select "Set image as background."
This is a crucial step because it forces Google's internal rendering engine to treat the image as a primary layer of the slide, which is essential for the conversion process in the next step.
STEP 3: Export to SVG
With your image locked as the background, go to File > Download and select Scalable Vector Graphics (.svg, current slide).
Google Slides will run its internal vectorization algorithm, which is surprisingly adept at handling soft gradients, and download the converted file to your machine.
You can now open this file in any browser or vector editor to see the sharp, scalable results.
STEP 4: Fix the Aspect Ratio Caveat
One minor issue with this method is that the exported SVG often defaults to a 1:1 (square) container, regardless of your initial page setup. This may result in empty white space or "gaps" on the sides of your image.
To fix this, open your new SVG in a free vector tool like Figma or Inkscape to quickly crop the canvas to the edges of your artwork. Even with this extra step, the gradient quality remains superior to most third-party raster-to-vector tools.
Want to keep unlock all the premium content?
Become a paying subscriber to get access to all other subscriber-only content.
✓ Full archive of premium guides with ready-to-use prompts
✓ Structured AI courses (step-by-step, start-to-finish)
✓ Every upcoming premium tutorial
Google is testing an agent in the Gemini app, internally named Remy, that can integrate with Google services to take actions on a user's behalf Google Chrome silently installs a ~4GB Gemini Nano model on desktop devices.
Anthropic researchers detail “model spec midtraining”, which adds a stage between pretraining and fine-tuning to improve generalization from alignment training.
Google updates AI Mode and AI Overviews to add “a preview of perspectives” from firsthand sources like social media and Reddit, presented as “Expert Advice”.
Google DeepMind takes a minority stake in the maker of Eve Online, a multiplayer role playing game set in outer space, and plans to train its tech on the game.
🌎 GPT5.5 Instant: OpenAI’s new default model for ChatGPT.
🎆 ChatGPT Images 2.0: OpenAI’s upgraded image generator
⚙️ Rork Max: AI tool to build native iOS apps
💻 Holo 3: Open AI agent that can use computers like a human
THAT’S IT FOR TODAY
Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!
See you tomorrow :)
- Dr. Alvaro Cintas