Is AI Creating New Value, or Just Rewriting Emails?

The current optimism in the semiconductor industry, which has led the Bank of Korea to raise its growth forecast, stems from the shift in computing demand. Unlike the previous digital economy driven by search and zero marginal cost reproduction, generative AI requires new computation for every query. Nvidia told investors that the top five hyperscalers are projected to pour $1.3 trillion into capital spending next year.

Despite this massive investment, a key concern is whether the computing power is creating new value that genuinely changes how people live, rather than merely making existing tasks like email writing or slide creation more efficient. The article posits that if AI primarily serves to automate or streamline old work, it risks forming an 'AI bubble' and an 'AI death trap' for companies.

The central challenge lies in transforming computing power into applications that yield novel solutions, such as drug discovery, instead of consuming it for tasks that could be handled by less powerful tools. This mirrors historical patterns where general-purpose technologies, like the steam engine, took decades to significantly impact productivity until industries reorganized around them.