Which jobs will not be replaced by artificial intelligence (AI)? Jobs that require complex decision-making and reasoning are difficult to replace in the short term. Kai-Fu Lee says that it is difficult to replace columnists for The New Yorker, but the same cannot be said for news compilation. Similarly, within the medical field, radiologists are more likely to be replaced compared to general practitioners. Jobs such as matchmaking transactions on internet systems and customer service for telecommunications providers are easily replaceable. This is evident from the significant layoffs at companies like British Telecom (BT) and Vodafone.
2021 was the year with the fastest increase in unicorns in history. However, the pressure of rising capital costs has led to a gradual decline in startup investments. According to a survey by US-based National Venture Capital Association (NVCA), venture capital investments amounted to US$171 billion in 2020, US$345 billion in 2021, and an estimated US$241 billion in 2022, a decrease of more than 30%. This is certainly related to the increase in capital costs. It was originally thought that the economic downturn would continue for a while, but it seems that AI is an important solution.
The key concept is to "provide personalized services based on vast amounts of data and individual needs." This concept can scale up to "small-scale diversified production" in factories and scale down to personalized everyday experiences. This will deepen the era of "winner takes all."
In the first stage of internet development, market formation occurred at a much faster pace than in the industrial age. However, typically, many startups would participate initially, and after a series of eliminations, the leaders would emerge. But now, users are actively involved in creating data, making data generation almost costless. This means that this market will experience "exponential growth," making it even more difficult for new startups to emerge.
However, major technology companies see great opportunities in AI. After the emergence of ChatGPT, many tech companies have launched related applications. The most influential one is still Microsoft, which has announced the integration of ChatGPT into all its application platforms. It was first deployed on Bing in February and then Office in March. Alibaba and Meta have also developed large-scale language models similar to GPT.
Furthermore, AI applications are ubiquitous, encompassing domains, images, text, and sound. The linkage and integration of multiple contents and hybrid applications are easily imaginable variations. This is the seamless convergence of diverse co-creation.
This trend has been discussed for several years, but it was unexpected that AI and Nvidia's GPUs would be the driving factors. Companies like TSMC and MediaTek have indirectly benefited from this. Additionally, Nvidia uses "proprietary software" to ensure that GPUs operate optimally, which is intimidating to other competitors.
From a user perspective, software makes AI expansion easier and more user-friendly. Many ecosystem players utilize various algorithms to provide effective applications through ChatGPT, and the advantages of the frontrunners are already evident. This is not just a short-lived hype.
Previously, Apple's App Store provided a platform for diverse applications, but OpenAI directly enters various applications, leading to even more astonishing diffusion of benefits. For example, GPT-4 can combine images, which is another application that surpasses everyone's imagination.
In the future, users will not simply search for partial information; instead, they will induce more relevant information from partial data. This revolutionary change prompts us to face a moment of redefining industries and national strategies.
