Welcome, AI Visionaries!
If you're watching where AI is actually headed in terms of not just generating text but thinking, moving, remembering, it’s time to look at the training grounds. Genie 3 gives agents a world. Hassabis says that the world is about to change. And Apple is quietly rebuilding the interface millions use every day to ask for answers.
📌 In today’s Generative AI Newsletter:
Genie 3 simulates full 3D environments for AI to learn in
Demis Hassabis says AGI will rival the Industrial Revolution
Apple’s AKI team aims to untether Siri from Google
GenAI.Works builds agent infrastructure for real-world deployment
Special highlight from our network
Teachers Save Hours Each Week with PLAUD
When time is scarce, the right AI tool makes the difference.
Educators today spend an average of 12 hours a week on prep and admin. PLAUD cuts that in half. The result? More energy for teaching, less stress, and better outcomes for students.
Structured Notes in Seconds so you can focus on teaching, not typing.
AI Summaries & Mind Maps proven to boost recall and exam performance.
Teach Smarter: PLAUD’s insights highlight where lessons connect—or where they don’t.
With PLAUD, classrooms run smarter.
DeepMind’s Genie 3 Could Be the Training Ground for AGI
Google DeepMind has unveiled Genie 3, a foundation world model that builds entire 3D environments in real time, not for passive viewing, but for AI agents to inhabit, interact with, and learn from. DeepMind sees it as a key milestone on the road to AGI, where agents no longer just respond to data but evolve through simulated experience.
Here’s what Genie 3 brings to life:
Simulates interactive 3D worlds using only a text prompt, generating photo-real or imagined environments at 24 fps, 720p, for several minutes at a time
Embeds memory into simulation, allowing the model to reason over long time horizons and maintain physical consistency as agents interact with objects
Supports promptable world events, enabling dynamic interventions like weather changes or object creation during a simulation
Builds on DeepMind’s Veo and earlier Genie models, combining physics-based video generation with real-time interactivity and open-ended learning environments
Designed to train general-purpose agents, helping them explore, adapt, and learn through trial and error inside endlessly generative virtual worlds
No, this isn't about making the video look better. Genie 3 turns AI from a viewer into a participant for real learning, a test bed where the agents of the future will stumble, recover, and eventually outgrow the lab. The public doesn't have access to Genie 3, but it is a big step forward in creating embodied AI agents.
📡Demis Hassabis: AI Will Be 10x Bigger Than the Industrial Revolution
Bloomberg / Getty Images
Demis Hassabis, the cofounder of DeepMind, says we’re nearing the final chapter before AGI changes everything. In a sweeping new interview, he reflects on his journey from chess prodigy to Nobel-winning AI researcher and outlines a vision of the future filled with radical abundance, medical breakthroughs, and a total rethinking of work, meaning, and purpose.
Here’s what Hassabis is seeing ahead:
Hassabis believes artificial general intelligence will emerge in five to ten years, exhibiting all human cognitive abilities. He says DeepMind remains focused on this “frontier research,” while others chase commercial gains.
Hassabis sees AI as a path to “radical abundance” if distributed fairly. “We’ll need philosophers and economists to help define what purpose means,” he says.
Hassabis wishes AI tools hadn’t gone public so fast. “If I’d had my way, we’d have left it in the lab longer,” he says, citing missed chances to cure diseases.
DeepMind staying in London was no accident. Hassabis resisted Silicon Valley’s pull, choosing to build British AI infrastructure and talent. Google’s new London HQ was a direct result.
DeepMind’s AI has beaten Go champions, mapped 200M proteins, and is now driving AI into every corner of Google. Hassabis says the next step is systems that reason, plan, and discover autonomously. The world may not be ready for AGI, but Hassabis is already living in the after.
🔗Apple's AKI Team: The In-House Effort to Untether Siri from Google
Photo by Justin Sullivan — Getty Images
Apple has formed a new team called AKI (Answers, Knowledge, and Information) with one job: turn Siri into a real-time, grounded answer engine that doesn’t need Google. Led by Robby Walker and reporting directly to Apple’s AI chief, the team is developing systems that crawl the web, retrieve facts, and deliver conversational responses all while staying private and native to iOS.
Here’s what’s coming together:
AKI is led by insiders. Former Siri lead Robby Walker heads the team under Apple AI chief John Giannandrea, signaling deep integration across the OS.
It goes beyond Siri. Apple is pursuing both a standalone chatbot app and embedded answer layers across core services, including Safari and Spotlight.
Search will stay private. Job listings point to systems that blend retrieval and ranking algorithms with Apple-grade privacy and on-device intelligence.
This time, it’s in-house. After trialing OpenAI integrations, Apple is shifting toward full-stack control of its answer engine, likely to reduce dependency on rivals
Apple’s current AI rollout has been met with yawns. AKI is its clearest move yet to reclaim the search moment on iPhones, as users drift toward conversational AI for fast, grounded answers. If the team hits its targets by spring 2026, Siri may no longer need Google to think for her.
GenAI.Works is building a platform where AI agents, regardless of vendor or stack, can connect, run tasks, and help real businesses grow.
Experts predict the AI agent market could grow to over $52 billion by 2030¹, and AI overall is expected to add $4.4 trillion in economic value each year².
Yet most people still don’t understand how these tools work or how to invest in them.³
That’s where we come in.
We’ve built the world’s largest and most engaged AI community, with over 12 million members and more than 2,500 tools, apps, and templates already in use.
Now, we’re building the technology that ties it all together:
The GenAI Protocol: helps AI agents work together.
The Distribution Engine: turns a simple prompt into a full business workflow, from copywriting to customer outreach.
Instead of trying to guess which AI startup will win, you can invest in the infrastructure they’ll all need to benefit.
Join us and be at the forefront of tomorrow
Reg CF offering via DealMaker Securities. Investing is risky.
¹ MarketsandMarkets, “AI agents market worth $52.62 billion by 2030,” PR Newswire, July 10, 2024.
² McKinsey & Company, “The economic potential of generative AI,” June 2023.
³ Pew Research Center, “AI in daily life,” March 2023.
🚀 Boost your business with us. Advertise where 12M+ AI leaders engage!
🌟 Sign up for the first AI Hub in the world.
