At GenInnov, we do not just invest in innovation businesses. We attempt to innovate ourselves. Our beliefs are shaped by our own experiences. The video at the link offers a glimpse of the suite of tools our small team has built in-house.

A Laboratory for Conviction

The video showcases only a fraction of what we have built. The true value of these tools lies not in their utility alone. They have been our laboratory for conviction.

When colleagues with no artistic instinct produced images that stopped us mid-scroll, when those who had never written code began shipping functional scripts, when junior team members delivered research of a depth that once demanded years of experience, and when legal filings took shape without a lawyer in the room, we needed no industry surveys. We were witnesses. The trajectory and velocity of improvement taught us to tune out the chorus cataloguing what AI could not yet do (Software: A Glass Half-Empty Analysis) and focus on what was rapidly becoming possible.

We discovered early that models do not need programmers to speak Gujarati. We watched Chinese models expound Kant in chaste Gujarati ourselves. We have saved substantial amounts in professional services. But the value runs deeper than cost. We would never have absorbed the intricacies of dozens of difficult subjects, let alone attempted global investments, without the debating tools we built internally. Experience, not theory, convinced us that allowing tools to roam across domains unlocks creative possibilities few anticipate.

From Experience to Conviction

These experiences shaped the arguments we have articulated in our notes. In one of two indulgent pieces on AI's epistemic underpinnings last year, we discussed why GenAI is not a killer app but app killers. Recent developments have made our other note equally relevant: why AI is a multitude, not a singularity. We explored this last weekend while examining the Moltbot phenomenon.

Anthropic's introduction of domain tools illuminates the boundaries of domain-specific AI. Domain boundaries matter for languages, law, medicine, and protein folding alike. Yet there must be limits to generalizability. Why should weather models blend with text-to-video? This is a subject we hope to explore over time.

We live in an era of instant copiability. Moats once carved in stone are now drawn in sand. We aspire to patent some of our work, and likely will. Yet we do not believe secrecy protects anything worth protecting. This is why we relish every opportunity to demo and discuss our tools with stakeholders. Anyone curious is welcome to reach out.

Where Advantage Now Lives

As the power of mere ideation wanes, competitive advantages are migrating. In old-fashioned terms, the last mile matters again. Personal history is a quiet advantage. The more we interact with AI systems that remember our queries and debates, the sharper they become for our needs. A virtuous cycle forms, one that will define the next battleground for AI companies themselves. Relationships, availability, and reputation have grown more important than ever, even in fund management.

Our tentative conclusions:

AI has tremendous utility. Those who disagree likely have not tried hard enough.

• AI cannot be imposed from above. It must become culture.
• AI is not merely about efficiency. It will birth new processes, products, and services.
• AI models should not operate alone. Their power multiplies when they debate and cross-check one another.
• Personal context is emerging as infrastructure. An AI that knows your history becomes irreplaceable.
• "Change is the only constant" now carries an entirely different weight.

We would not mind if AI progress halted tomorrow, whether due to scaling limits our historians have prophesied or regulations yet to come. The disruption has already cleaved the world into two camps: those who attempt and those who sit. We are brimming with ideas. Given time and capital, what this video showcases will soon feel unremarkable. The tools we proudly demonstrated early last year barely merit a footnote here.

Until stagnation arrives, we choose to keep attempting. For all our stakeholders.

GenInnov's consultants, Swaraj Gambhir and Sidharth Mishra, have extensively worked on the production of the video and all the tools highlighted in the video