Discover more from Life in Color
 AI Apps: Going Beyond the Hype
[AI Apps Usage Data, Skeuomorphic Design]
Sharing an essay from our Evolving Internet Insights Newsletter unpacking some AI Usage Data and how AI might be stuck in its skeuomorphic phase.
(P.S. For the summary and further breakdown of this essay, follow me on Twitter)
Note: Some sections of this essay were originally released on my Evolving Internet Insights newsletter.
At the onset of new technological paradigms, people often ask, “how real is this technology?” in their attempt to rationalize their skepticism of something new.
We saw this happen in the numerous Web3 boom and bust cycles that have taken place in the last decade. During the peak of a cycle, there is usually a lot of attention and hype for the new technology in question. The headlines and media coverage over embellish what is actually happening on the ground as it relates to progress. At the peak of the last Web3 cycle, it seemed like everyone was talking about Web3 values, its use cases, and how it was going to change the world.
Of course like most cycles, that came and went.
Since the launch of ChatGPT-4 in March 2023, the AI fervor has now taken the world by storm. Today, we are experiencing the same kind of cyclical, Web3-like euphoria. It seems like every business is now trying to become an “AI” business and every startup pitch deck has some mention of AI.
This begs the question: how real is the hype?
AI Usage By The Numbers
ChatGPT is the fastest application ever to reach 100M users, achieving this coveted milestone in just two months. To put this into context, it took Instagram 30 months and TikTok 9 months to hit the same milestone.
ChatGPT crossing the 100M user mark in just two months is an impressive feat no matter how you slice and dice it. But it is also important to go deeper than the top of the funnel metrics. For instance, are users returning to the app and using it regularly in their workflow?
A leading VC firm, Sequoia, wrote a blog post that double-clicked on this topic. They found that while the initial growth was impressive, the retention rate for “AI-first” companies was lower than that of social media platforms and leading consumer apps. The median retention rate for AI-first companies is 42% compared to a median retention rate of 63% for existing social and consumer apps.
The report also goes on to highlight that user engagement on AI platforms is also significantly lower. Using DAUs/MAUs (daily active users as a percentage of monthly active users), a commonly used indicator to assess how often users are going back to the app every day, the median for AI platforms is 14% compared to 51% of the social media and consumer platforms. Using this metric, WhatsApp is best-in-class with 85% DAUs/MAUs.
All of these data points seem to suggest that AI applications initially garner a lot of consumer hype, but then the novelty of using these AI tools fizzles quickly. Anecdotally, we are observing a similar trend within our network of companies of all sizes and geographies. People sign up for these applications, feel the “wow” factor from how well AI applications like ChatGPT work, and then get stuck finding new use cases.
From a use case perspective, we believe AI is currently in its skeuomorphic phase.
The Skeuomorphic Phase
Skeuomorphism is a design philosophy that imitates the form and appearance of real-world objects, creating a sense of familiarity and comfort for users. The new thing retains elements of the old thing it’s trying to replace.
By giving users recognizable elements to engage with on the surface while subtly introducing them to new concepts and technologies in the background, skeuomorphism serves as a bridge between the old and the new – encouraging user to adopt / use emerging technologies.
Skeuomorphic design is everywhere.
For example, early iterations of the smartphone interface (for Apple's iOS and Google's Android operating systems) utilized skeuomorphic design elements to help users transition from traditional tools to touch-based smartphones that could do “it all.” The calculator app is a great example of how designers wanted to mimic the look of an actual calculator in its smartphone-based cousin. Similarly, the notes app looks like an actual notebook, and the news app looks like an actual bookshelf.
Using this framework, AI seems to be in the skeuomorphic phase as it relates to adoption and use cases.
The chart below highlights that a few of the most common words used in ChatGPT prompts are “write” and “create” with the next most popular set of words (“list,” “explain,” and “research”) treating ChatGPT effectively as a search engine. It seems many users are using AI technologies like ChatGPT for the use cases that are most obvious to them: in this case, using ChatGPT to help us write and find information. While being able to carry out these tasks with ChatGPT is impressive in its own right, we are just scratching the surface. It will be up to companies like OpenAI to help users go beyond the skeuomorphic phase.
Confusing the Appetizer for the Entree
All fundamental technologies went through their own version of users gravitating towards initial use cases that were obvious to them. But we eventually broke out of the skeuomorphic phase because we realized the underlying technology enables much more than the initial use cases.
When mobile phones were first introduced, if we had just stayed in the skeuomorphic phase, we would have just considered mobile phones as “just phones in our pockets.” Of course, they became much more than that. Mobile phones became a new computing paradigm.
The mobile phone ended up becoming a de facto computing interface that impacted all sectors of the economy. For example, the rise of social media and improvements in smartphone cameras allowed anyone with a smartphone to become a content creator overnight.
For AI, it enters its post-skeuomorphic phase when users stop seeing AI just as an intelligent copywriter, research assistant, or a “search engine 2.0” that facilitates the jobs we need to get done, and instead start seeing it as an intelligent computing paradigm. For example, an emergent use case for AI is AI companions, which is one of the fastest growing categories in Gen AI. Companions represent a best-in-class form of personal assistants (read: Siri, Alexa, Google but with emotion baked in) that are more personal and nuanced and can be accessed by the user through conversation.
Early use cases that are easy to understand (skeuomorphic) bring in early users. Early adoption brings in more builders to build products for a growing user base which ideally drives more adoption for the emerging technology.
This is all great for initial adoption.
But it is important to not confuse the appetizer for the entree.
👇 Get with the times, subscribe to Life in Color 🥳
🙏 Share Life in Color with a friend and ask them to subscribe
🙏 Share on Twitter / Linkedin with a short note
🙏 Share on your company Slack / Teams channels and communities
This post is provided for educational and informational purposes only. Nothing written in this post should be taken as financial advice or advice of any kind. The author(s) may own some of the NFTs, art and/or collectibles mentioned in this post. The content of this post are the opinions of the authors and not representative of other parties.
Empower yourself, DYOR (do your own research).