After the Buzz of Multimodal Models – What’s Next on the Path to Superhuman Intelligence?

I just learnt that multimodal AI might already be Passé.

Building models that absorb vast quantities of human language and knowledge was an incredible leap. Yet it reminds me of a pivotal AlphaGo moment - when the AI initially mirrored human strategies before eventually abandoning them entirely for ideas no human had conceived. [Watch AlphaGo Documentary]

That’s the real frontier we’re heading towards.

As David Silver from DeepMind recently put it, we are now entering the "Era of Experience" - where AI no longer simply learns from human data, but builds its own understanding through direct interaction with the world.

Imagine AI discovering entirely new treatments in medicine.

We can already glimpse this future. xAI’s Grok, for instance, is starting to interact freely on X, creating its own threads of experience. But maybe the true "AlphaZero moment" for the real world will only arrive when robotics matures enough for AI to explore the physical world without relying on human datasets at all. Who knows......

Beyond human knowledge. Beyond human ceilings.

Most Enterprises Are Still Only at the Beginning

Despite the excitement, enterprise adoption of AI is still in its infancy:

• 42% of companies have tried to deploy AI in at least one function.
• Yet AI impacts just 5 to 10% of overall business operations today.

The implication is clear: while pioneers are surging ahead, most organisations remain tentative, missing early opportunities to reshape their future.

David Silver’s recent commentary is a warning: models purely trained on human data are effective for automation, but stifle innovation. To thrive, businesses must now embrace AI systems that learn, adapt, and discover independently.

After the explosion of AI innovation following ChatGPT’s rise, I’ll be honest — even I haven't fully figured out how I should personally pivot and adapt in this new world. The pace of change is relentless, and it sometimes feels like just keeping up is an achievement in itself.

As an individual, adapting to this new AI-driven world is challenging. Yet paradoxically, I believe it may be easier for enterprises. Businesses today have a unique opportunity to harness AI tools and drive real outcomes, particularly by collaborating with companies like Oracle investing heavily in AI research and business innovation.

At Oracle, we want to help organisations move beyond automation into true experience - driven AI. Here’s where you can start:

🔹 Harness Existing Data - But Plan Beyond It Use Oracle Cloud’s AI services to extract insights from your enterprise data - but also invest in models that can continuously adapt as new patterns emerge.

🔹 Embed AI into Real-Time Operations With Oracle Fusion Applications and Autonomous Database, bring AI into the heart of daily decisions - from supply chain to finance - ensuring your business continuously learns and improves.

🔹 Focus on Self-Learning Systems Take advantage of Oracle’s Machine Learning services on OCI to develop models that retrain themselves using live operational data - eliminating the need for constant manual rework.

🔹 Prepare for a Post-Human Data Era Stay close to Oracle’s Research and Development initiatives in reinforcement learning and self-supervised learning - critical foundations for staying competitive as AI evolves beyond human-derived knowledge.

I’m still navigating what this new AI era means for me personally. What about you? How are you thinking about adapting yourself, not just your business?