Google just destroyed the Competition ... | Ep. 40 🏄‍♀️

Welcome back to the Innovation Network Newsletter

This week felt like watching two parallel realities collide. In one reality, Google just dropped Gemini 3 Pro and obliterated every benchmark, nearly doubling Arc AGI 2 scores while OpenAI scrambles with an internal panic project called "Shalotpeat".

In another reality, we're raising a generation so coddled by curling parenting and fear-mongering media that young people need guidance on how to start conversations at networking events.

Patrick and Aragorn spent nearly two hours unpacking everything from why Google's comeback matters, to whether AI-generated art can ever truly be "art," to why we desperately need philosophers embedded in every organization building AI systems.

Also: Matthew McConaughey and Michael Caine just licensed their voices to 11 Labs, and Dutch artists are creating anti-surveillance masks.

Let's dive in.

Patrick, Aragorn & Nikola

We're Curling Parenting a Generation Into Incompetence

Patrick opened with something that's been nagging at him. He recently did a workshop for young bankers at AO Shearman, around 25 years old, early in their careers. Before the session, the organization held a prep meeting to teach these professionals how to have meaningful conversations at networking events.

Then he discovered a Swedish/Norwegian company called Wing People that literally provides companions to networking events for people afraid to start conversations on their own.

His kids struggle with the same thing. So he started researching what's actually happening.

The "slow life strategy" data is stark: 

Today's 22-23 year-olds are less likely to drive, work, date, or drink alcohol than previous generations were at 17-18. Adolescence is stretching longer. Young adults are taking more time to develop resilience, grit, and basic social skills.

Aragorn pushed back on the easy answer:

The deeper problem

A Dutch orthopedic specialist posted about how the Netherlands' Minister of Education tweeted that "proper educational progress requires comfort." The specialist called this a massive misconception that's been scientifically debunked many times.

We're seeing the rise of a culture specifically in the West where everything is about comfort and convenience. And it's actively degrading human development, both physically and mentally.

Aragorn added: "I got a Commodore 64 when I was seven. I spent loads of time on my computer and internet. I still developed into a human being who can start a conversation and pick up a phone."

Why does it matter 
We're actively creating a generation with limited resilience, weak social skills, and extended adolescence just as we're entering an era that demands adaptability, rapid learning, and constant reinvention. Technology isn't the only problem. Screens aren't destroying kids. Helicopter parenting, risk-averse culture, and media fear-mongering is an overlooked aspect!

Gemini 3 Pro Scares Open AI

Google just released Gemini 3 Pro and it's not hyperbole to say they've leapfrogged everyone. Nearly doubled the Arc AGI 2 benchmark score. Dominated across almost every leaderboard. Made everyone who said "Google is finished" look foolish.

The Technical Leap

Arc AGI 2 is the benchmark that matters most right now. It tests AI's ability to understand abstract conceptual patterns, from colored shapes and forms where you identify patterns and predict outcomes. For humans, it's relatively easy. For AI, it's been hard. Until now.

Gemini 3 Pro made a massive 15% jump over GPT-5 Pro. Gemini 3 Deep Think more than doubled the score compared to previous top models.

The Strategic Focus: Spatial Intelligence

Looking at the LM Arena leaderboards across different categories—text, web dev, vision, text-to-image—Gemini Pro is dominating almost everything except search.

The consensus? Google doubled down on making Gemini capable of understanding the physical world through vision.

Patrick demonstrated by asking Gemini to create a virtual gallery of Dalí's artworks. One sentence prompt. The result? A complete timeline, full gallery, copyright-compliant images, and even a "talk to Dalí" chat integration.

Aragorn showed another feature most people don't know exists: Guided Learning mode in Gemini. If you want your kids to learn, let them go at it with Guided Learning. This is the perfect tutor in AI that helps you do your homework in a way that helps you develop the large language model up here in your skull.

Imagen 3 (Nano Banana Pro): Visual Reasoning Unleashed

Patrick tested Nano Banana Pro extensively. It's not just making "cool images" anymore. It generates text correctly, creates diagrams, animations, dashboards, infographics, even handwriting.

He showed an infographic created from the Innovation Network URL in one prompt. Then asked for a Star Wars version. Instant transformation with correct text and design.

The comic book example was jaw-dropping: He fed Gemini the first chapter of his book about audio wave evolution. One prompt: "Make a comic out of it." Et voila!

Why does it matter
Google's focus on spatial intelligence and visual reasoning represents the next evolution of AI. Models that understand the physical world, not just language. When AI can look at your screen, understand what you're doing, see through your camera, and help you interact with reality, we're no longer talking about chatbots. We're talking about true AI assistants that bridge digital and physical worlds.

Identity as a Service

Matthew McConaughey and Michael Caine just licensed their voices to 11 Labs. McConaughey is deep in tech. He's building his own LLM, advising Mark Benioff at Salesforce, and invested in Elevenlabs. This isn't a celebrity doing a cash grab; it's someone who understands the technology making a strategic bet.

Why does it matter
We're watching the beginning of digital immortality. not in the sci-fi sense, but in practical, commercial reality. When McConaughey and Caine license their voices now, they're creating assets that outlive them. When game studios pay voice actors to train AI models, they're not stealing. On the contrary, they're creating a new licensing category.

What Makes AI-Generated Content "Art"?

Patrick stumbled into one of philosophy's best questions while looking at AI-generated artwork. He found research by Daniel Parris showing how people feel about AI in different aspects of art creation.

The findings? People already accept AI for some tasks like translating subtitles, generating visual effects. It's a cascading model of acceptance. But with a limit!

So Patrick asked the harder question: "Can AI truly create Art? Or is art a very human thing?"

Aragorn suggested we might be headed toward redefining what "art" means. Potentially distinguishing between art made by machines and art made by humans, valuing them differently.

There is an ambiguous, nebulous thing, a lack of understanding and knowledge about how something came to be. And that's exactly that part of our brains which we don't fully understand. It's exactly that part in AI that we can't understand and can't see, which is on the one hand the challenge and the fear of AI right now. And on the other hand, that's also the magic. That's where the magic resides. That's where the seat of the soul sits.

Why does it matter
How will we value human creativity in an age where machines can produce technically perfect work at scale? If we redefine "art" to require human creation, we preserve a space for human meaning-making even as AI matches or exceeds technical execution. If we don't, we risk commodifying creativity into pure technical output. The deeper insight is that ineffable human experience of creation and interpretation. That ambiguity, that lack of perfect understanding of process. AI can replicate the output, but it can't (yet) replicate the context of suffering, joy, discovery, and humanity that gives art its soul.

We Desperately Need Philosophers in Every Organization

Patrick was struck by an interview with Amanda Askell, a philosopher at Anthropic, on 60 Minutes. If you as an organization are thinking about deploying AI and maybe even building your own large language model, then the character you want it to behave, the constitution you want to give that AI, can only be derived from good conversations with a philosophical edge.

Aragorn agreed completely, but with caveats: "We should start to really consider what it means to be a philosopher. A true philosopher is somebody that adheres to the rules and the laws of epistemology".

We're living in a world where that's increasingly not the case because people are increasingly looking at postmodernist thinking and saying 'whatever is real for you might not be real for me, maybe we're all living in our own realities.' These kind of lines of thought may be entertained but should never be taken too seriously because it basically breaks us apart.

In reality, AI is forcing us in so many ways to look at ourselves. What is our place in the world? Who are we? Where do science and the things we haven't been able to explain yet meet? What does that mean for us?

Testing AI as Philosopher

Patrick decided to test this. He fed Claude the trolley problem. The classic ethical dilemma about whether to switch a train track to kill one person instead of five.

Claude gave an extensive answer using utilitarian approach (maximizing overall good). But Patrick pushed further with the variant: What if you have to push a fat person onto the tracks to stop the train? Claude didn’t hesitate:

Why does it matter
We're building AI systems that will make millions of decisions affecting billions of people, and most organizations are approaching this as a technical problem rather than a philosophical one. The trolley problem is literally being coded into self-driving cars, medical triage systems, and resource allocation algorithms. We need polymath thinkers who understand both the technology and the centuries of philosophical work on ethics, epistemology, and human values. Otherwise, we're building the most powerful decision-making systems in history on a foundation of unexamined assumptions.

The Future of Gaming: AI NPCs & Infinite Worlds

Three major developments converged this week around AI and gaming, painting a picture of where interactive entertainment is headed.

SEMA 2: Your AI Gaming Coach

Google released SEMA 2: an AI agent that watches you play games and helps you understand the environment. It's like Microsoft Copilot for gaming (which already exists on new Microsoft laptops with NPU chips for games like Minecraft).

World Labs: Fei-Fei Li's Spatial Intelligence Play

World Labs (founded by AI pioneer Fei-Fei Li) is creating 3D environment generation. This connects to the broader pattern Patrick identified: "First we created a picture which becomes high resolution. A video is nothing else than 60 pictures in a row. And now we're building entire world models."

Aragorn is particularly excited about Fei-Fei Li: "She's of a generation quite a bit younger than Geoffrey Hinton and John LeCun... I'm very excited about what she wants to do and what she wants to bring to the world of gaming."

The promise? "Infinite, downloadable content for any game that you would like to have."

Why does it matter 
We're watching three technologies converge: AI companions that understand gameplay (SEMA), AI that generates infinite explorable worlds (World Labs, Google Genie 3), and AI NPCs that behave intelligently and naturally. When these merge the distinction between "playing a game" and "living in an alternate reality" becomes meaningless. You'll have AI guides helping you navigate infinite AI-generated worlds populated by AI beings that seem alive.

Anti-Surveillance Masks Hit the Streets

Dutch artist Jip van Leeuwenstein created masks that make it impossible for AI to recognize you through facial surveillance.

This is literally straight from Kai-Fu Lee's 2041 predictions. He wrote about how AI's rise would lead to fashion trends including masks to avoid AI surveillance.

Jeff Bezos' Project Prometheus

Jeff Bezos is launching Project Prometheus. A $6 billion robotics venture focused on embodied intelligence, transport, and logistics.

The name references the Greek god who stole fire and gave it to humanity (with consequences). Fitting for a project that might reshape manufacturing.

The strategic angle

This connects to everything discussed about spatial intelligence and world models. Those 3D AI-generated environments aren't just for gaming—they're training grounds for robots. Simulate millions of scenarios, train robots in virtual worlds, then deploy to physical reality.

Jeff Bezos is the man behind the biggest robotization and automation move in the entire world. Amazon has more robots than employees.

But there's another dimension. Bezos has been heard saying he believes in data centers in space. Combined with Blue Origin (his space company) and this robotics push, he's positioning for a future where AI and robots operate in space—where humans are vulnerable but machines thrive.

Join Us: December 11th Meetup

Innovation Network Meetup - December 11th in Amsterdam at AIM

You'll meet robotics experts, investors, people who know vibe coding. Discussion tables on everything covered in this episode and more. No brand pitches, no product demos—just authentic sharing and learning.

Newsletter full video recording