Key Takeaways:
– Doubt clouds the potential of GenAI as industry analysts question the return on investment.
– Issues like lack of a definitive application, data availability, chip shortages, and power problems challenge GenAI expansion.
– Goldman Sachs voices concern over GenAI’s path, comparing it to past failed tech hypes like 5G and blockchain.
– MIT Professor Daron Acemoglu estimates that GenAI will automate only 5% of tasks within a decade.
– On the other hand, advocates like Joseph Briggs and Kash Rangan at Goldman Sachs maintain a positive outlook on GenAI’s future.
As the world watched GenAI blossom impressively over the past year and a half, it inevitably invited criticism. Industry analysts are currently doubting whether the significant investments poured into the technology will yield desired returns. These concerns were expressed in a Goldman Sachs Research letter that highlighted not just the lack of a “killer app,” but issues such as data availability, chip shortages, and power challenges that might impede GenAI’s growth.
Historical Iterations of Digital Hype
This isn’t the first time tech advancements have garnered mixed reviews. Tech journalists, who have witnessed technological booms and busts such as the dot-com disaster, the rise of cloud computing, and the proliferation of smartphones couldn’t help but compare. Big data, the tech obsession of the early 2010s, experienced a similar rollercoaster ride before crashing in 2019. After this, the big data marketing machinery began exploring new territories. From blockchain to 5G and self-driving cars, several other trends attempted to grab attention and funding but fell short.
Enter GenAI
The situation changed when OpenAI launched a new large language model (LLM) called ChatGPT in late 2022, thereby putting neural network-based GenAI on the radar. The hype for GenAI, promising to be a game-changer, left analysts questioning its fate.
Notable tech moments have gone two ways in the past. Some moments proved to be considerable turning points, like mobile and cloud technology, while others such as the dot-com boom and Hadoop-style computing took years to reveal their full impact. GenAI’s trajectory is thus uncertain, prompting questions about its position in the future.
The Goldman Sachs Stance
Painting a less than optimistic picture of GenAI, Goldman Sachs issued a report titled “Gen AI: too much spend, too little benefit?”. The editor Alison Nathan expressed concern over the absence of concrete results despite heavy spending in infrastructure.
Daron Acemoglu, MIT professor, who participated in the report added that only a quarter of tasks that AI aims to automate would feasibly be automated cost-effectively within a decade. Data quality could also pose an issue if ramped-up production alone was relied upon to accelerate GenAI’s progress.
The chip shortage for training GenAI models could further obstruct progress. Apart from this, the steep costs of training and using GenAI could outweigh the productivity or efficiency benefits it might deliver. Power requirements for training LLMs and other GenAI models could also become a significant concern.
Positive Outlook On GenAI
But not all opinions on GenAI’s future are pessimistic. Joseph Briggs, a senior global economist at Goldman Sachs, projected that GenAI could automate 25% of all work tasks and boost US productivity by 9% over the next decade. He suggested that the cost of technology tends to decrease rapidly over time, which could trigger an explosion of GenAI applications and consequently, higher automation rates. A similar view was echoed by Kash Rangan, and Eric Sheridan, who remain optimistic about GenAI’s long-term potential.
With the clock ticking, the question looms large – will GenAI produce significant returns before time runs out, or go down as another failed tech hype? The answer, only time will tell.
