Daron Acemoglu, Institute Professor at MIT, suggests that only around 25% of tasks suitable for artificial intelligence (AI) are expected to be cost-effective in the next decade. Despite potential breakthroughs, the impact of AI is not expected to be immediate. According to Goldman Sachs, this means that AI will likely affect less than 5% of all tasks, contributing only a 0. 5% increase in US productivity and a cumulative 0. 9% growth in GDP over the next decade. Acemoglu stated that while large language models have proven impressive, it remains uncertain whether these models alone can match the intelligence of Hal from "2001: A Space Odyssey".

He expressed concerns about the limitations of the current architecture and suggested that increasing GPU capacity may not accelerate AI progress as much as improving the quality of data. Acemoglu also questioned the source of the necessary data. Jim Covello, Head of Global Equity Research at Goldman Sachs, stated that AI needs to solve complex problems in order for companies like Nvidia, Microsoft, Alphabet, Meta, Amazon, and Super Microcomputer to achieve a return on the $1 trillion expected to be spent on AI capital expenditures in the coming years. Covello emphasized that there is currently no cost-effective use for AI and highlighted a significant misunderstanding about the technology's capabilities, stating that it lacks cognitive reasoning abilities.