What Big Tech Missed And How Startups Can Still Win
Insightful talk on AI entrepreneurship and innovation strategies.
FOR WHOStartup founders
DenseTalkExpert
FOR WHOStartup founders
DenseTalkExpert
FOR WHOStartup founders
DenseTalkExpert
Context
Alexandre LeBrun, CEO of AMI, discusses his entrepreneurial journey, including selling Wit.ai to Facebook and his current venture, Amie Labs, which focuses on developing world models for AI.
Key points
Alexandre LeBrun shares his experience of receiving an email from Mark Zuckerberg, which he initially thought was spam, leading to the acquisition of Wit.ai by Facebook. 1:40
LeBrun explains the concept of world models, which learn directly from the real world through sensory data, unlike large language models that rely on text. 8:33
He discusses the challenges and costs associated with building foundational AI models, including the need for significant GPU resources. 6:00
LeBrun emphasizes the importance of being ambitious and taking risks, especially in the European startup ecosystem, which he feels lacks ambition. 24:32
He advises early-stage founders to focus on solving narrow problems while maintaining a large vision for the future. 27:27
LeBrun describes the potential applications of world models in robotics, allowing robots to operate in open environments safely and effectively. 9:32
He highlights the limitations of large language models in achieving general intelligence and the need for direct world experience in AI models. 12:19
LeBrun recounts his experience at Facebook, where he learned the value of ambition and the willingness to take bold steps. 26:00
He discusses the strategic decision to base Amie Labs in Paris and other cities instead of San Francisco, emphasizing commitment and focus. 23:24
LeBrun envisions a future where AI with common sense can significantly improve the quality of work and life. 28:38
Quotes
The real cost of this 1.2 billion was not dilution. The real cost is expectations.
A machine can speak about a cat, but if the machine has never seen a cat, it will never be very smart.
If you choose a very narrow problem, have a very big vision.