SuperAI 2025
"AI progress is overestimated in a year, but underestimated in 10."
Those words still echo in my mind two weeks after SuperAI 2025.
This conference brought together thought leaders, innovators, and builders to reflect on where AI stands today—and where it's heading.
Key Takeaways
- AI is here to stay, but AI is not ready, and may never be ready to make critical decisions—there are things that still require a human element
- To build a defensible company you gotta build vertical
- Thin ChatGPT wrappers somehow work?
I've seen some crazy ways AI is applied like cell cytometry, robotics simulation, sourcing of manufacturers for physical products (Razer gaming assistant).
But I've also learnt of its dangers—it's not very observable, it lacks the human factor in critical decisions and it makes us complacent.
Standout Moments
-
Manus AI shared their fascinating product journey, highlighting not just their pivots, but the why behind them. A masterclass in adaptability and vision.
-
A gripping panel on AI in Bio-Sciences, featuring Munjal Shah, Alex Aliper, and Mahyar Salek, explored how AI is reshaping biology—from discovery to commercialisation—and the nuanced balance between innovation and ethical responsibility.
-
Edward Snowden delivered a sobering reminder: As AI systems scale, their biases become harder to detect—yet their impact on human lives grows. What's more concerning is no single person or organisation is accountable for these consequences. Final decisions must remain in human hands, and AI systems must be made more transparent and observable.
Beyond the sessions, I met a diverse and driven community of young founders and AI enthusiasts—each exploring how AI could revolutionise their industries. Their energy and curiosity was contagious.
I'm incredibly grateful to NYC and Wavesparks for the opportunity to participate in this transformative experience. Let's stay grounded and skeptical, yet optimistic and innovative to the applications of AI in our industries and daily lives.