- Future Tools
- Posts
- Bezos is BACK
Bezos is BACK
Microsoft + Nvidia + Anthropic = 🤝

Welcome back! Anthropic CEO Dario Amodei just went on 60 Minutes and said the quiet part out loud: AI could wipe out half of all entry-level white-collar jobs within five years. He believes AI will cure cancers, outperform humans at most cognitive tasks, and—if left unchecked—cause social damage on the scale of cigarettes or opioids.
It’s a striking message from the leader of one of the world’s most powerful AI labs.


Jeff Bezos Heads Up New AI Startup

Via TechCrunch | Image Credits: Emma McIntyre/WireImage
Bezos is back in the arena. This time, he’s taking an operational role as co-CEO of Project Prometheus, a new AI startup that’s already raised $6.2 billion.
What Prometheus is building: The startup is focused on “AI for the physical economy”—tools that can accelerate engineering and manufacturing across industries like aerospace, automotive, and advanced computing. The vision resembles companies like Periodic Labs, which simulate physical processes to train AI models.
Who’s involved: Bezos is partnering with Vik Bajaj, who helped lead Google’s life sciences division, co-founded Verily, and has deep experience in biotech and applied research. The startup already has nearly 100 employees, including talent from Meta, OpenAI, and Google DeepMind.
Why it matters: Bezos is betting that the next wave of AI will connect directly to the physical world. We’re rapidly moving beyond just chatbots to creating systems that help build real products.
Microsoft and Nvidia Pour $15B Into Anthropic
The biggest AI alliances are getting even tighter. Microsoft and Nvidia just announced a pair of investments totaling up to $15 billion into Anthropic—a move that pushes the company’s valuation toward $350 billion and deepens the three-way partnership across chips, cloud, and model development.
Key commitments from the deal:
Anthropic will purchase $30 billion of Azure compute, plus contract up to 1 gigawatt of additional capacity.
Nvidia and Anthropic are forming their first deep engineering partnership, co-designing models and optimizing Nvidia architectures for Claude.
Anthropic will also commit up to 1 gigawatt of Nvidia’s Grace Blackwell and Vera Rubin systems.
Microsoft will broaden access to Claude models across Azure, Foundry, and the Copilot ecosystem.
What it means for Anthropic: The startup now becomes the only frontier model available across all three major cloud providers—AWS (its primary training partner), Google Cloud, and Microsoft Azure—giving it unmatched distribution and compute optionality.
The bigger picture: These deals underscore the increasingly interdependent nature of the AI ecosystem. As Microsoft CEO Satya Nadella put it, AI is now too big and too complex for “zero-sum narratives or winner-take-all hype.” The companies building the future of AI are tying their infrastructure, chips, and models together.
Google Launches AI-Powered Travel Planning Inside Search
Google just turned travel planning into a one-prompt experience. A new feature inside AI Mode lets users describe an upcoming trip and instantly generate a full itinerary inside Google’s Canvas workspace.
Current features:
Auto-build itineraries with flight options, hotels, activities, and reviews
Refine results with follow-up prompts (cheaper hotels, activities within walking distance, alternate dates, etc.)
Pull photos, location data, and ratings from Google Maps
What’s next: Google is expanding agentic booking, allowing AI Mode to help users book restaurants, events, flights, and hotels through partners like Booking.com, Marriott, Resy, StubHub, and many more.
Why it matters: Travel planning is becoming an AI-first workflow, right inside the world’s most-used search engine. And as agents begin to book things for you, Google moves one step closer to being a full-service digital concierge.


From prompt to presentation instantly

Via PageOn.Ai
PageOn.AI turns prompts, PDFs, documents, or mind maps into polished, interactive presentations or infographics using autonomous AI agents. It handles research, design, layout, and data citation automatically.
How you can use it:
Convert long text into visual slides or graphics
Generate charts, diagrams, and media blocks automatically
Edit everything using a Notion-style drag-and-drop interface
Pricing: free and paid plans

Offline AI note-taking with superpowers

Via Submind
Submind is an AI-powered, privacy-focused note-taking app that works fully offline. It indexes all your notes, summarizes them, transcribes voice recordings, and embeds media—with all data stored locally.
How you can use it:
Search across all notes instantly
Summarize dense content offline
Record audio and get automatic transcripts
Store multimedia notes with local-only privacy
Pricing: free and paid plans available


Jobs, announcements, and big ideas
Google introduces Gemini 3, its most intelligent model to date.
Manus releases a browser extension that automates workflows across CRMs and logged-in platforms.
Anthropic’s Claude models roll out to Microsoft Foundry and the Microsoft 365 Copilot preview.
Intuit signs a $100M+ partnership with OpenAI to bring TurboTax, QuickBooks, and Mailchimp into ChatGPT.
Amazon’s Zoox begins offering free robotaxi rides in San Francisco to challenge Waymo.
Google launches Scholar Labs to explore how generative AI can support advanced research problems.
xAI launches Grok 4.1 across grok.com, X, and its iOS and Android apps.
NVIDIA debuts Apollo, an open family of AI models for scientific simulation adopted by major industry partners.


Can Gemini 3 dethrone ChatGPT? I put it through the hardest tests I’ve got ➡️

That’s a wrap! See you Friday.
—Matt (FutureTools.io)