Amid the rapid growth of AI applications, Andrew Ng, founder of DeepLearning.AI and Managing General Partner at AI Fund, emphasizes that Generative AI (GenAI) encompasses much more than just chatbots, showcasing a diverse array of practical scenarios where the new technology can be applied.
Ng acknowledges that while users may be experiencing fatigue from chatbot applications, there remains significant potential for product enhancement and utilization across various sectors.
Connection to Taiwan
He advises businesses to pinpoint industry-specific AI use cases before deciding whether to develop solutions internally or outsource them. Ng’s connections with Taiwan are substantial, involving collaborations with Foxconn and an engineering team for DeepLearning.AI based there. In May, Taiwan’s National Science and Technology Council announced a collaboration with Ng’s AI Fund on a venture capital project, potentially through joint ventures or capital investments.
Ng shared his insights during the AI Fund’s recent online forum, where he engaged in a conversation with Roy Bahat, head of Bloomberg Beta. Although Ng joined Amazon’s board in April, he did not represent Amazon during the forum.
Big-league competition
When discussing the current state of AI development, Ng noted that AI is increasingly capable of performing a broader range of tasks, presenting opportunities for growth and associated risks.
In response to Sequoia Capital’s article “AI’s $600B Question,” Ng explained that the article focused on elements like model training, GPU procurement, and capital investment. He reassured that unless companies are directly competing with model training providers like OpenAI, Anthropic, or Google, they need not be overly concerned.
He further explained that the significant investments made by these major players benefit the overall ecosystem by reducing costs for application developers.
Build or buy?
For companies, Ng stresses the importance of clarifying their industry-specific use cases to decide whether to build or buy AI solutions. He suggests that knowledge workers should utilize GenAI tools to enhance productivity, with companies providing the necessary training for their employees.
Ng states, “Individuals skilled in using AI will outperform those who are not.” However, he points out that this transition will not happen overnight, and some risks might be overstated. He adds that specific industries, such as finance or healthcare, typically have inherent risk management mechanisms for AI applications.
Chatbot fatigue: more applications out there
Ng also mentioned that tools offered by companies like OpenAI have lowered the barriers to developing chatbots, resulting in an abundance of such applications and some user fatigue. Nevertheless, there are numerous other consumer application areas to explore. For example, payment service providers can integrate AI functions into their backend systems to categorize expenses into entertainment, equipment, or other categories, thereby improving product usability.
A few years ago, developing an application could take up to six months, including market research and product development, with concerns about recouping development costs. With significantly lowered development barriers, developers can release multiple applications simultaneously and test market reactions, creating substantial value with even a few successes.
Ng also highlighted the emerging trend of agentic workflows or AI agents, which automate longer processes from single commands. In document writing, for example, AI assistants can draft, research, revise, and identify issues within the document, suggesting improvements and making revisions. This comprehensive approach, compared to issuing single commands to a language model, can significantly enhance user productivity.