#321 Nick Frosst: Why Cohere Is Betting on Enterprise AI, Not AGI - Eye On A.I. Recap
Podcast: Eye On A.I.
Published: 2026-02-17
Duration: 1 hr 1 min
Summary
In this episode, Nick Frosst, co-founder of Cohere, discusses why the company is focusing on practical enterprise AI solutions rather than the ambitious goal of AGI. He emphasizes the importance of transformer architecture and how Cohere aims to make it useful for businesses.
What Happened
Nick Frosst, a co-founder of Cohere and former researcher at Google Brain under Jeff Hinton, shares his journey into the world of neural networks, starting from his undergraduate days at the University of Toronto. He recalls being introduced to neural networks and the excitement surrounding their potential, particularly around the time of AlexNet's groundbreaking paper in 2012. Frosst emphasizes that while he initially felt he missed the boat on AI, he has remained engaged with developments in the field since then.
The conversation shifts to Cohere's mission, which is distinct from many others in the AI landscape that are focused on achieving AGI. Instead, Cohere is dedicated to solving practical problems in the enterprise sector. Frosst explains that while transformer architectures, which power many current AI models, have been around since 2017, the focus for Cohere is on making these models scalable and useful for businesses. He highlights the company's efforts to leverage transformers in a way that meets the specific needs of enterprise clients, rather than chasing the more theoretical goals associated with AGI.
Key Insights
- Cohere focuses on practical enterprise AI applications rather than AGI.
- Transformers remain the primary architecture for AI models, with ongoing innovations in scaling and training.
- Nick Frosst's background at Google Brain influenced his perspective on AI's development.
- The ambiguity of AGI's definition continues to shape the conversation in AI research.
Key Questions Answered
What is Cohere's focus in the AI landscape?
Cohere's focus is on practical enterprise AI applications, steering clear of the more theoretical goal of AGI. As Nick Frosst mentions, the company aims to solve real-world problems within businesses by making transformers useful for their needs, rather than pursuing the broader and often ambiguous objectives associated with AGI.
How did Nick Frosst get involved with AI?
Nick Frosst's journey into AI began during his undergraduate studies at the University of Toronto, where he was introduced to neural networks. He worked with Jeff Hinton at Google Brain, which significantly shaped his understanding and approach to AI, particularly in foundational research and its applications.
What are transformers, and why are they important?
Transformers are a type of architecture that has been pivotal in the development of AI models since their introduction in 2017. Nick Frosst points out that while there have been various enhancements and scaling techniques, the core architecture remains the same. This stability and scalability is what makes transformers particularly useful for creating effective language models.
What distinguishes Cohere from other AI companies?
Cohere distinguishes itself by prioritizing practical AI solutions for enterprises instead of pursuing AGI. As Frosst explains, many companies are still grappling with the concept of AGI, but Cohere aims to deliver usable AI tools that can be effectively integrated into business operations.
What challenges does Cohere face in the AI market?
Cohere's main challenge lies in navigating a competitive AI landscape that is heavily focused on AGI. Frosst acknowledges the ambiguity surrounding AGI and stresses that Cohere's commitment to solving enterprise-level problems may set it apart but also requires clear communication about their goals and the value they offer.