Anticipating the Generative AI Hype Cycle

Goldman Sachs recently released a report titled "Gen AI: Too Much Spend, Too Little Benefit?" (direct link to report PDF), offering a particularly myopic view of the benefits of generative AI in light of the tremendous capital that has been invested in chips, training, infrastructure, and human talent. As is the case with many flashy headlines, I think it's worth digging into the details, and in this case, scrolling through the PDF to evaluate some of the opinions it contains. Here's one:

 

Jim Covello: "I place low odds on AI-related revenue expansion because I don't think the technology is, or will likely be, smart enough to make employees smarter."

 

Quite a hot take, and one I see as patently absurd (one must question Goldman's motives here!) given the acceleration we have already achieved with clients and internally. However, I find it plausible and even a bit of a relief to see that Goldman is identifying the proverbial "peak of inflated expectations" in the hype cycle.

 

 

If we've learned anything about implementing generative AI solutions so far, it's that there's a lot of work to do, in both strategy and implementation. The implementation is not just about prompting a model but in providing them with good data, process/workflow, coordination of multiple agents, human interface, infrastructure and more. The best solutions not only will have a human in the loop but will give that human incredible leverage to be more effective, intelligent, and efficient.

 

I do see the report as a bellwether and if anything very positive news that greater understanding reasonable expectations - not just hype - are just around the corner. Now it's time for the rubber to meet the road as we'll see more stories of successful acceleration. As usual, startups will lead the way with enterprise incumbents following.

There's never been a better time to build. Let's go!