ChatGPT is well known to be the fastest app to reach 100 million MAUs (monthly active users) — and it did so organically in just 6 weeks. By comparison, it took Instagram 2.5 years, WhatsApp 3.5 years, and YouTube and Facebook 4 years to reach that level of user demand. But ChatGPT is not an isolated phenomenon. Character AI’s deep engagement (average session time of 2 hours), Github Copilot’s productivity gains (up 55%), and Midjourney’s path to profitability (hundreds of millions of dollars in revenue) all indicate that the first killer apps are here.
Developers are key. Developer-first companies like Stripe or Unity have realized that developer access opens up use cases you can’t even imagine. In the past few quarters, we have received proposals ranging from music generation communities to AI matchmakers to AI customer support agents.
The modalities are evolving. The first versions of AI applications were mostly autocomplete and first draft, but these modalities are now becoming more sophisticated. The camera pan and fill features south korea mobile database introduced by Midjourney are a great example of how generative AI-first user experiences are becoming richer. Overall, the modalities are evolving from personal productivity to system-level productivity, from humans in the loop to execution-focused agent systems.
Copyright, ethics, and existential fear. The debate on these sensitive topics continues. Artists, writers, and musicians are divided, with some creators rightfully angry that others are profiting from derivative works, while others embrace the new AI reality (think Grimes’ profit-sharing proposal and James Buckhouse’s optimism about becoming part of the creative genome). No startup wants to be the Napster or Limewire of the ultimate Spotify (thanks, Jason Boehmig). The rules are still unclear: Japan has declared that there are no IP rights to content used to train AI, while Europe has proposed strict regulations.