As GenInnov begins its journey as a licensed firm, we explore why true disruption lies not in AI per se, but in machines solving what humans cannot

Towards the World of Ever-Improving, Super-Einsteins

At GenInnov, we believe the early 2020s marked humanity’s discovery of how to manufacture intelligence. First came the Chatbot Era, where AI dazzled us like a cheeky butler from a Wodehouse novel, chatting, joking, and answering trivia. Cute, but basic. Now, in the Agentic Era, we’re over mere talk. We want AI to book flights, fix selfies, tame spreadsheets, and maybe prank-call our pals—little helpers buzzing on our phones, still trapped in their digital cribs.

The truly exciting developments, however, lie just over the horizon. The imminent third stage is where intelligence breaks free from its silicon confines. No longer tethered to technological gadgets, it begins to permeate everyday objects. Whether you call it automation, robotics, embodied AI, or simply the "era of smart everything," the trajectory is clear: our physical world is about to become cognitively enhanced, capable of adapting and learning from experience. This transformation is fueled by action models, experience models, multi-modal capabilities, and hardware innovations designed to amplify cognitive function.

Yet, even this is a mere prelude. The fourth and ultimate stage will dwarf all that came before: an era where humanity's most intractable, long-standing, and complex problems are tackled with unprecedented cognitive prowess, far surpassing the limits of even the most brilliant human minds. This post-AGI era is closer than many believe. Already, some AI models are rivaling prodigious high school math talents, scoring on par with gold medalists in Olympiads. Soon, they will exceed the capabilities of the greatest human geniuses across every discipline.

The clearest and earliest harbingers of this new age of innovation—envision millions of tirelessly improving super-Einsteins at our disposal—will undoubtedly emerge from healthcare and drug discovery. It is here, more than anywhere else, that we will first witness the immense power and transformative potential of generative AI fully unleashed.

Stage 4:The First Proofs are Here from Smarter Diagnostics 

The third and fourth stages of AI—where robots get smart and super-Einsteins solve cosmic riddles—will shake up the world way more than today’s chatty bots and phone helpers. Everyone’s buzzing about robotics: self-driving cars, machines that run marathons, fold laundry, or even brainy vacuums are stealing the show. But the truly jaw-dropping transformation is happening elsewhere—less visible, more profound. This is when AI starts cracking humanity’s toughest puzzles at a level we can’t even imagine. That’s why GenInnov isn’t just an AI fund—it’s an innovation fund, betting on what we call the Innovation Era, not just an AI Era.

Take a peek at the table below—AI’s making diagnostics a breeze! From X-rays to cancer scans, it’s speeding things up and catching problems with awesome precision. This isn’t just helping doctors—it’s showing AI can dig deeper into complex problems than any human brain.

This diagnostic superpower is just the start. It proves machines can tackle questions at a level of complexity we’ve never reached, even with our best experts. And if AI can do this for health scans, imagine what’s next for fields like drug discovery, where the stakes are even higher, and the puzzles even tougher.

Footnote: Our compilation of studies is not necessarily FDA-approved or peer-reviewed, to showcase the rapid pace of improvement. For emphasis, most of these claims are unverified and approximate. The table is presented to highlight how new intelligent machines are surpassing previous methods in diagnostic fields, tackling complexity that exceeds human capability. Even if much of the above is eventually proven exaggerated or unimplementable, the improvements are likely to continue at a breakneck pace.

The Bigger Test—Drug Discovery

Diagnostics showed us AI’s knack for spotting patterns, like a super-smart detective reading X-rays. But AI’s real brainpower shines in drug discovery, where it’s not just looking but thinking through mega-tough problems to invent new medicines.

Until now, our understanding of biology has been shaped by the limits of human cognition—we’ve used broad classification systems, grouped proteins and pathways, and relied on approximations across biochemical space just to make sense of the complexity. These frameworks helped us get this far, but they’re crude by comparison. Machines aren’t bound by those constraints. They can process molecular relationships, disease signatures, and mechanistic pathways at a depth and dimensionality we could never hold in our minds.

The next level is reasoning, hypothesizing through data systematization that goes beyond what we, with our current tools, have been able to do so far, generating new ideas, and solving complex scientific problems.

Unlike diagnostics, where the task is to classify or detect, drug discovery demands creativity, iteration, and long timelines. The problems are harder. The unknowns are greater. And until recently, most of the work was slow, manual, and staggeringly expensive. But that’s changing.

The simplified table below offers only a taste of what’s to come. While it tracks visible improvements, the real story lies in the quality and breadth of possibilities now unlocked—massively more experimental pathways, novel molecular architectures, and predictive models that operate at scales and complexities unthinkable before. For example, models like Absci’s zero-shot antibody generators can now propose viable therapeutic structures without prior examples—something that defies how humans typically reason through protein engineering. Likewise, Recursion's phenomics platform enables high-dimensional screening across tens of thousands of compound-cell combinations—an effort that would take years by traditional methods.

AI’s Visible Impact on Drug Discovery Pipelines

This Is Only the Beginning

The horizon stretches far beyond these early signs. The table above is a postcard from the future, not the full picture. The true transformation is not about faster molecule generation or shorter timelines—it’s about the explosion of possibilities now available. We are witnessing the emergence of entirely new ways to interrogate biology and chemistry, from generating de novo protein structures with functional constraints to predicting full phenotypic outcomes from single-molecule perturbations.

Think of what's unfolding across different domains:

• Synthetic biology is advancing with models that design complex gene circuits that work on the first try.
• Antibody engineering is being reshaped with zero-shot generation techniques.
• Molecular docking and binding affinity prediction are leaping ahead with diffusion and transformer models.
• Pathway simulation is moving toward whole-cell and even organism-scale modeling.

Each of these advances is not merely additive—they’re multiplicative. They change the shape of what’s possible.

Of the Innovation Singularity

Much of the public’s fascination with the AI era still revolves around familiar and almost theatrical examples—machines that write articles like this one for marketing pitches, compose music, pass tests, or book travel itineraries without being asked. These are clever, even delightful, but they belong to a relatively trivial class of progress. Most of the focus today is on which model is best to work on one’s private data pool or how many data centers it takes to power the next chatbot.

A growing number, however, now grasp the deeper transformation underway: the automation of thinking itself in real-world environments. This is the rise of embodied AI, robotics, and autonomous systems that don't just respond—they act. As these thinking machines begin to drive forklifts, manage warehouses, and inspect power grids, we’ll enter an era of global productivity unlike anything before.

But it’s the stage beyond that that will most radically reshape our world.

And this is where the bigger idea takes hold. Words like AGI or superintelligence are only placeholders—our way of describing something we’re sensing but don’t yet fully grasp. The more useful idea is this: the ceiling on human cognition, which has remained stubbornly fixed, is finally being lifted. For centuries, we extended our reach through better instruments. Now, we are extending our reach through better minds—ones we’ve built.

We’re not just building faster tools—we’re welcoming entirely new minds into the process of discovery. Minds with unfamiliar strengths, strange reasoning, and perspectives we might never have imagined. The risks are real, as are the disruptions, but the arrival of this intelligence explosion is no longer a question of if.