# Is Cohere the Next AI Powerhouse? | First Time Founders with Ed Elson Source: https://www.youtube.com/watch?v=JqGAQMYwuzQ Recap page: https://rapidrecap.app/video/JqGAQMYwuzQ Generated: 2026-03-01T13:34:06.653+00:00 --- ## Quick Overview Cohere, founded by former Google Brain researchers, focuses on building foundational large language models (LLMs) specifically for the enterprise market, differentiating itself from consumer-focused AI by prioritizing security, private data access, and tailored solutions for businesses like Dell, SAP, and Salesforce, which is a strategic move given the high cost and resource intensity of building foundational models from scratch. **Key Points:** - Cohere focuses exclusively on developing foundational AI models for the enterprise market, contrasting with consumer-facing AI. - The company was founded in 2019 by three former Google Brain engineers: Nick Frosst (featured speaker), Aidan Gomez, and Ivan Zhang. - Foundational models like Cohere's are resource-intensive, requiring massive compute power and data, making them difficult for most companies to build internally. - Cohere's unique value proposition is providing secure, private access to these powerful LLMs through an agentic platform, allowing enterprises to use the models on their own data. - The development of large language models, particularly Transformers, is seen as a pivotal moment in AI history, similar in impact to the Industrial Revolution. - Host Ed Elson questioned the potential for AI development to exacerbate wealth inequality, but Cohere's founder suggested that while the technology is transformative, the focus should be on building useful tools for businesses rather than solely pursuing Artificial General Intelligence (AGI). - Cohere has secured significant funding, reportedly raising over $100 million this year alone, reflecting strong investor confidence in the enterprise AI sector. ![Screenshot at 00:05: The title card for the podcast episode, 'First Time Founders with Ed Elson', introduces the interview format and the guest, Nick Frosst, co-founder of Cohere.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-00-05.jpg) **Context:** This video is an interview between Ed Elson, the host of the "First Time Founders" podcast, and Nick Frosst, the co-founder of Cohere. Cohere is a company specializing in building and deploying large language models (LLMs) for enterprise clients, contrasting with models designed for general consumer use. The discussion centers on Cohere's enterprise focus, the technical barriers to entry in foundational model development, the geopolitical implications of AI advancement, and the company's strategy for serving businesses securely. ## Detailed Analysis The conversation opens by establishing Cohere's focus on building foundational AI models specifically for the enterprise sector, contrasting this with general consumer AI. Nick Frosst, co-founder, explains that building these foundational models is incredibly resource-intensive, comparing it to building a rocket. Cohere's strategy is to serve large enterprises like Dell, SAP, and Salesforce by allowing them to use these powerful models securely and privately on their own data via an agentic platform, which is a key differentiator from consumer offerings. Frosst notes that only about ten companies globally possess the resources to build these foundational models currently. The discussion shifts to the historical context of AI, noting that the advent of Transformer architecture marked a significant shift, analogous to past industrial revolutions, making AI development highly relevant to the economy. Host Ed Elson raises concerns about AI potentially exacerbating wealth inequality. Frosst acknowledges this concern, suggesting that policy and ensuring broad access to beneficial AI tools are crucial to prevent an extreme divide. He contrasts the current landscape with earlier AI efforts where models were trained narrowly (like for chess), explaining that modern LLMs trained on the entire web offer much broader utility but also present greater challenges in terms of control and deployment. Frosst concludes that his focus is on creating broadly useful, powerful technology for enterprises rather than chasing the speculative goal of Artificial General Intelligence (AGI). He also shares his personal background, mentioning that he studied under Geoffrey Hinton at the University of Toronto, which heavily influenced his research focus on neural networks and their potential. ### Cohere's Enterprise Focus - Building foundational AI models for enterprises (Dell, SAP, Salesforce) - Focusing on security and private data access - Differentiating from consumer-facing AI. ### The AI Landscape - Foundational model development requires immense resources, akin to building a rocket - Only about 10 companies can currently build these models. ### AI's Impact and Geopolitics - The Transformer architecture is a transformative technology - AI development could influence geopolitical power balance, but Cohere focuses on practical enterprise utility rather than AGI. ### Challenges and Philosophy - The risk of AI increasing wealth inequality is a major concern - Cohere aims to build useful tools that augment human work, not replace it entirely. ### Founder's Background - Nick Frosst studied under Geoffrey Hinton at the University of Toronto, influencing his focus on neural nets and foundational models. ### Future Outlook - The technology is expected to continue rapidly augmenting and automating tasks across organizations. ![Screenshot at 00:05: Title card for 'First Time Founders with Ed Elson' featuring Nick Frosst.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-00-05.jpg) ![Screenshot at 00:20: Nick Frosst discusses how AI is becoming embedded in how the world operates, mentioning major enterprise clients like Dell, SAP, and Salesforce.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-00-20.jpg) ![Screenshot at 01:13: Ed Elson asks Nick Frosst how he got involved with Cohere, mentioning co-founders Aidan Gomez and Ivan Zhang.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-01-13.jpg) ![Screenshot at 03:37: Nick Frosst explains that the cost and resources required for foundational models are a significant barrier to entry.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-03-37.jpg) ![Screenshot at 27:25: Advertisement slide for LinkedIn, sponsor of the podcast segment.](https://ss.rapidrecap.app/screens/JqGAQMYwuzQ/00-27-25.jpg)