Meta wants eyes

Plus: Amazon’s agentic bet

Welcome back! Zoom meeting coming up on your calendar? Soon, you can show up as an AI clone of yourself. Upload a photo, and Zoom will generate a photorealistic avatar that mimics your movements in real time—perfect if you rolled out of bed two minutes before a call. 

It’s one more step toward Zoom’s CEO’s vision of a “digital twin” that takes meetings and emails on your behalf. Not quite there yet, but the gap between you and your AI stand-in is getting really small.

Meta Connect 2025 Puts Wearables at the Center

Mark Zuckerberg at Meta Connect 2025 / Via: Meta

Meta just wrapped its Connect 2025 event, where Mark Zuckerberg took the stage to unveil a lineup that makes it crystal clear: Meta wants AI on your face.

  • The headliner: the $799 Meta Ray-Ban Display. At first glance, they look like chunky Ray-Bans. But pinch your fingers and a screen appears in your right lens—messages, maps, Instagram Reels, all floating in your field of vision. It’s the closest we’ve come to what Google Glass promised a decade ago.

  • For athletes: the $499 Oakley Meta Vanguard. Wraparound design, sweat resistance, built for people who run, ride, or train outdoors. Think less “tech demo” and more “gear you actually want to wear.”

  • For everyday use: the $379 Ray-Ban Meta Gen 2. 8 hours of battery life, sharper 3K video recording than the previous gen, and a charging case that holds two days of power. These are the glasses that finally feel practical.

Meta isn’t stopping at hardware. Developers are getting a new Wearable Device Access Toolkit, letting apps tap into the glasses’ vision and audio sensors.

Big picture: While the metaverse is still in the mix, Connect 2025 signals a pivot in Meta’s priorities. The company is betting on AI wearables as the next computing platform—small enough to replace some phone habits and powerful enough to open an ecosystem for developers.

Amazon Debuts Agentic AI for Sellers

Amazon is bringing a new always-on AI agent to third-party sellers—turning its Seller Assistant into more than a helper.

What it does: The upgraded agent handles tasks on its own, from inventory monitoring to compliance checks. It flags slow-moving products, suggests pricing tweaks, and ensures listings meet safety regulations across markets. Sellers remain in control, but much of the manual work is automated.

Example use case:

  • Spotting products that risk storage fees and recommending actions.

  • Preparing shipping strategies based on demand forecasts.

  • Auto-checking compliance rules in different countries.

Why it matters: This is agentic AI in action—systems that don’t just chat but actually do. Amazon is betting on a future where commerce runs on autonomous agents, able to buy, sell, and strategize with minimal human input.

DeepMind’s Gemini 2.5 Wins Programming Gold

Google DeepMind says its latest model hit a milestone: It’s the first AI to win a gold medal at the International Collegiate Programming Contest.

What happened: Gemini 2.5 solved a real-world problem that stumped human programmers—figuring out how to move liquid through a complex duct system—in under 30 minutes. Out of 12 tasks, it nailed 10, ranking second among 139 of the world’s best student coders.

The big picture: Programming contests require creativity, reasoning, and working code. DeepMind’s win suggests AI is moving from beating humans in games like chess and Go to tackling open-ended scientific and engineering problems. For Google, it’s more evidence that Gemini isn’t just a chatbot, but a research tool with real-world problem-solving power.

Days of Manual Effort Completed in Minutes

Most AI helps with work. Notion AI finishes it. Your Notion Agent doesn’t just suggest what should be done—it completes tasks end-to-end, exactly how you like them.

Your Agent can:

  • Audit and update your knowledge base

  • Wrangle scattered customer feedback

  • Draft project updates and docs

  • Fine-tune your product roadmap

  • Critique and rewrite comms

Ready to delegate to your Notion Agent?

Browse Smarter Not Harder

Genspark

Genspark weaves AI into the browser itself, with the option to run models locally for privacy.  Its autopilot mode automates repetitive clicks, while integrations connect across 700+ apps.

How you can use it:

  • Summarize long YouTube videos instantly

  • Automate repetitive web tasks without coding

  • Compare products and prices across sites

Pricing: Free

Your AI Business Partner

Budera

Budera acts like a virtual growth team for entrepreneurs. It scans markets, tracks competitors, and delivers daily ROI-driven tasks through its “Growth GPS.” Its Trend Hawk feature spots industry shifts, while Competitor Radar keeps tabs on rivals.

How you can use it:

  • Get daily recommendations to grow sales and revenue

  • Track competitor activity in real time

  • Identify early signals of industry changes

  • Ensure compliance with built-in legal checks

Pricing: Free & paid

Jobs, announcements, and big ideas

  • Luma AI teams with Adobe on Ray3, a generative video model built to take on Google.

  • Notion 3.0 brings AI Agents that can build pages, databases, and automate workflows.

  • NVIDIA and Intel join forces on AI infrastructure and next-gen PCs.

  • Microsoft reveals plans to spend billions on AI data centers.

  • Google Gemini app now lets users share custom Gems.

  • Google Chrome rolls out new AI tools for safer, smarter browsing.

  • Streamlabs launches AI streaming co-host with Inworld and Nvidia.

AI in the kitchen? Meet Zippy, the robot chef serving Michelin-star dishes.

That’s a wrap! See you next week.

—Matt (FutureTools.io)

P.S. This newsletter is 100% written by a human. Okay, maybe 96%.