There is a massive AI overhang today.
The term is analogous to snow overhang. When snow slides down a roof, sometimes a section goes over the edge and is seemingly supported by nothing. You know it will eventually fall; the tension is in the air. It's a question of time.
I have the same feeling with applying AI to products today. Sure, lots of people are talking about the potential of ChatGPT, but the difference between what is practical and possible, versus what is actually done today, remains large. Hence the overhang.
What's going on? A couple of hypotheses:
- There is no overhang; there simply are not many use cases.
- There are many use cases but companies don't have good ideas for how to effectively apply it.
- Most product teams understand what to do but haven't had time to react (a cycle is often 3 months, and reaction time 2-4 cycles).
- PMs know what to do but are in a holding pattern: the space is moving so fast technically, and product ideas and applications discovered so quickly, that it is easier to wait, let someone else take the risk, and follow fast.
Out of the above, I mostly believe (2) and (3).
The second, because the categories of Prompt engineering, Context engineering, LLM Product, etc are so fresh that only the tech enthusiasts (myself included) have done significant experimentation.
The third, because even great ideas will be deprioritized relative to well-understood high-impact work. Anecdotally, I've heard from a friend working on AI at a scale-up that the release of ChatGPT lit a fire under the business and product teams, and it got much easier to greenlight projects.
Since ChatGPT was released in November last year, the AI features entered the roadmap in Q1'23 at the earliest -- which is probably why we're seeing announcements right now for AI integration into products like Slack, ChatSpot, Discord, Notion, and probably countless more. Q1 is ending; if the OKR was to ship then time is running out.
Related: Transformative AI overhang (2020)