Aidan Gomez, the CEO and co-founder of Cohere, a Toronto-based startup, played a significant role in the development of the transformer architecture, which has been crucial in the advancement of generative AI models like ChatGPT. In a recent interview with The Associated Press, Gomez discussed the importance of transformers in building large language models and the focus on scalability and production readiness at Cohere.

One example of how Cohere is utilizing AI models is in the healthcare space, where doctors can use passive listening devices to pre-populate patient notes, saving time and increasing efficiency. Addressing concerns about AI models being prone to errors and bias, Gomez highlighted the use of retrieval-augmented generation models like Command R, which allow for fact-checking and reduce hallucinations.

Gomez also addressed public misconceptions about generative AI, emphasizing that the technology has the potential for profound positive impact and that developers are actively working to mitigate risks. While major technology companies are pursuing artificial general intelligence (AGI), Gomez stated that his focus is on maximizing the effectiveness of current AI technology for real-world applications.

Looking ahead, Gomez highlighted the potential for AI models to adopt external tools and exhibit more agent-like behavior, enabling them to interact with new software programs or APIs effectively. This shift towards more versatile and practical applications of AI technology could lead to significant advancements in various industries.