The Gartner Hype Cycle will be remembered as one of Gartner’s most notable achievements. It has been referenced countless times in settings ranging from sales and marketing pitches to board rooms; part of a universal language for the lifecycle of technological innovation. The Gartner Hype Cycle, once a cornerstone in the lexicon of technology strategists and marketers, has died. It was killed by the blinding pace of innovation in Artificial intelligence, more specifically, Natural Language Processing.
Smart phones followed a similar path on the Hype Cycle; introduced in the mid-2000s, they enjoyed a rapid ascent to the Peak of Inflated Expectations, fueled by media frenzy and consumer excitement. Over the next decade, as challenges were addressed (remember ‘you’re holding it wrong’?) and capabilities expanded, these devices moved into the Plateau of Productivity, becoming an indispensable part of daily life.
In contrast, the pace of innovation in the AI/NLP/LLM space has been unprecedented, and that’s an understatement. 2023 alone witnessed major advancements from companies like OpenAI, which released both GPT-3.5 and GPT-4, Midjourney released versions 4, 5, and 6 in the same year, each offering stunning improvements, the release of Google’s Bard, Anthropic’s Claude, PerplexityAI and too many others to list here. For context, as of the time of this writing, Hugging Face has 513,380 models posted on their site, each with different specialities, parameter counts, capabilities, and use cases.
This past week was no different. We were just getting our heads around short videos of 4–10 seconds when OpenAI released SORA, which allows for one-minute videos of never-before-seen quality and Google announced Gemini 1.5 pro, which allows for a context window of 1,000,000, tokens (increasing the common token limits by more than 5X). Humanity has never experienced this pace of innovation.
All of this mayhem means that the Hype Cycle is no longer suitable for measuring the adoption of technology in an AI-enabled world. The moment one technology comes close to peaking, a new groundbreaking advancement is released, perpetually holding the market in a state of never-ending Peak of Inflated Expectations. This phenomenon transforms the market adoption curve from a peak with a steep decline and a trough, into a series of minor disturbances, trivializing the once-significant process of technological maturity and market acceptance.
In addition to the breakneck speed of development in AI and NLP, there’s a growing cultural acceptance of technological imperfection. The nostalgic reliability of past technologies, like the consistent dial tone, is a stark contrast to today’s frequent and readily accepted technological glitches (every online meeting still begins with ‘can you hear me ok?’). This shift diminishes the presence of the Trough of Disillusionment, as society’s tolerance for imperfection rises and the continuous stream of innovations feeds its craving for newness and advancement (dopamine hit, anyone?).
I recognize that some may believe that the hype cycle is alive and well, and that this is just a much bigger market, creating the tallest ever Peak of Inflated Expectations. However, the undeniable utility and transformative power of these technologies would render any potential Trough of Disillusionment so shallow as to be almost imperceptible. As important, it is now clear that the barrier to entry for new companies in the AI/NLM space is sufficiently low that solutions for every vertical market, and every size organization, are going to dramatically reduce the time to achieve truly useful products and services.
The demise of the Hype Cycle is emblematic of the profound changes ushered in by Natural Language Processing and other AI technologies. The way we live and work is undergoing a seismic shift. These changes, while disorienting and challenging for some individuals and organizations, cannot be ignored without risking obsolescence. Markets are evolving and being reshaped at a pace unprecedented in human history, and the ability to navigate these rapid transformations will determine which organizations survive, and which ones don’t.