Coming down from the peak of the Generative AI hype cycle, and where we go from here.
FM (Friday Morning) Reflection #4
We’re finally coming off the peak of the first hype cycle with Generative AI, and I couldn’t be more thrilled. It’s time to do less talking and more doing with this once-in-a-generation leap in technology.
How do I know we’re coming off the peak of the hype cycle? There has been a recent uptick in the media about lackluster profits in GenAI businesses, leaders in the chip industry losing their footing, corporate managers fretting about the long road to ROI, and how the impact of GenAI has been overblown.
Here’s the reality: AI isn’t going anywhere.
If you look at where investors are putting their future bets – in sectors like utilities and energy, for example — AI is clearly reshaping demand and the market. The New York Times reported yesterday that utilities were the 2nd best performing S&P 500 sector so far this year, gaining 15.5%. Energy companies are busy inking agreements with Google, Microsoft, and Amazon to provide new sources of energy to power data centers, which may lead to a resurgence of nuclear power generation in the U.S.
Anyone who writes off Generative AI as just another passing trend will be in for a surprise. Believe me: This isn’t the end. It’s just the beginning — and the start of making GenAI useful to everyday people like you and me.
This shift is happening right now, quietly, behind the scenes in the way we work, without shouting it from the rooftops. I see people every day finding little shortcuts — ways to save a few minutes here and there to streamline their workday. Maybe they’re not drawing attention to it, because their company’s IT policies haven’t caught up, or maybe they’re worried about what others might think if they found out that they were using AI in their work. But they are.
People are voting with their feet – or this case – with their fingers by adopting GenAI tools into their daily routine.
Think about this: Generative AI, and specifically the large language models we keep hearing about, are giant repositories of knowledge, fact, opinion, philosophy, and prose. They’ve absorbed a massive amount of what we’ve written, discussed, or debated throughout history, and we now have access to that in a way where we can ask questions and receive a point of view – or more accurately stated – a synthesis of points of view, that we can use in formulating a useful conclusion.
The benefits of this can be huge. You know that old saying, “If you have a question, you should ask, because someone else is probably wondering the same thing”?
With these AI models, not only has someone likely asked the question that’s on your mind, but they’ve also likely found and published an answer—and now that answer is available to you.
AI is not a magic wand, and it’s up to us to develop the wisdom to use it responsibly.
We should be critical in judging the output of models, just like we do when we evaluate any single source of information:
Does the output make sense? Can I corroborate it with other authoritative sources? How did I ask the question? Did I ask the model to explain its reasoning, and whether it has confidence in its answers?
That’s why we need active engagement from all types of users to figure out what works, how to keep people safe, and where we can find the most value.
And it also means we need to prepare our society to responsibly use GenAI.
This will require elevating our expectations for education, stepping up our game on media and data literacy. We need to expand the teaching of critical thinking and ethics, foster higher-order decision-making skills, and we must find ways to teach math and science in ways that are engaging and relevant.
If we do these things, and as we come down from the peak of inflated expectations, we can bend the curve. We can shorten the dip in the trough of disillusionment to reach the slope of enlightenment faster, and from there, make a shorter climb to the plateau of productivity that is to come with a broad, ethical, and responsible adoption of Generative AI.
Less talking. More doing. Let’s get going.
Bonus Content
Reflection courtesy of the rain and storms buffeting the Mid-Atlantic region this morning from post-tropical cyclone Debby.
Cloud-obscured sunrise as I wrote this reflection today. Enjoy!