The AI Hype Cycle in 2024

While they’re not always right about timing, @Gartner_inc has done a pretty good job of forecasting sentiment with their ‘hype cycle’. Hypothesis about AI: we’re at the Peak of Inflated Expectations and we’re headed for turbulence.

Innovation Trigger (2023)

chatGPT is generally released and it’s stunning and then somehow even more stunning from v3 to v3.5 to v4. You don’t need a lot of money or expertise to start using it and it’s amazing.

Peak of Inflated Expectations (2024)

Where does this gravy train? Does it? Investors fund startups that say they’ll disrupt the messy work of humans figuring out what they want to do and AI will do everything.

Trough of Disillusionment (2025)

It turns out that generative AI is reductive and hyper-parasitic: it’s a calculator for telling you what you want to hear. Everyone’s producing the same thing and it’s of little real value to them outside of rote tasks.

Slope of Enlightenment (2025/2026)

Actually, we weren’t at the end of the road with the current gen. of LLM’s, but your product (or process) has to do something interesting, either giving the AI better context or better data (Khan Academy's 'Khanmigo' is a good example).

Plateau of Progress (2026+)

This is beyond me. Possibly it’s just more of the above. Possibly, computer vision and physical automation speed up and we get some real Jetson’s level stuff.

How do you beat the curve? Well, I work a lot with AI- it’s the department I’m in at UVA Darden and I teach in a data science program (MSBA). I also use a lot of genAI at my investee startup (Jedburgh- gamified language learning).?

I see just 1 thing over and over again: you have to figure out what the valuable intervention is and then work backwards to leverage AI. That’s how you go beyond the easy and obvious. It’s also what Hypothesis-Driven Development is about! https://hdd.works/aihype