AI companies have been attracting significant investment and gaining high valuations, leading many to question whether the industry is experiencing a bubble. However, Nick Frosst, co-founder of Cohere, a company that builds custom AI models for enterprise customers, disagrees with this view. He believes that while there may be some froth in the market, calling it a bubble undermines the genuinely useful features and automation that AI companies like Cohere are providing to their customers.
Frosst emphasizes the tangible value that AI technology brings, citing instances where their models have enabled new features or automated processes that were previously cumbersome. He argues that such practical applications of AI make it difficult to label the industry as being in a complete bubble.
However, Frosst does not share the same optimism as some of his AI peers when it comes to the development of artificial general intelligence (AGI), which refers to human-level intelligence. He believes that achieving AGI will take a long time, if it ever happens at all. This perspective contrasts with the views of figures like Mark Zuckerberg and Jensen Huang, who are more bullish on the potential of AGI.
Frosst acknowledges that AI technology is incredibly powerful and useful, but he cautions against viewing it as a “digital God.” He believes that more people are realizing the limitations of AI and adjusting their expectations accordingly. At Cohere, they strive to be realistic about what AI can and cannot do, focusing on the types of neural networks that can provide the most value.
Cohere’s business model is based on the research work of their co-founder and CEO, Aidan Gomez, who made significant contributions to the development of AI models like the transformer model. According to Frosst, Cohere utilizes a large language model as a foundation to create custom models for their enterprise clients. They believe that while specialization is important, a solid understanding of language in general is crucial.
Despite their belief in larger, foundational models, Frosst does not think that companies should rely on a single model to handle all tasks, whether consumer-oriented or B2B. He advises companies to focus on specific areas and have a clear understanding of what AI technology can and cannot accomplish.
Overall, Frosst and Cohere maintain a realistic perspective on AI technology, recognizing its immense value while avoiding extreme rhetoric about its capabilities. They believe in the power of AI to deliver significant benefits but do not see it as a threat to humanity. This balanced approach allows them to navigate the AI industry with sobriety and focus on providing genuine value to their customers.