Interview, Fireside Chat, Podcast
The State of Consumer Tech in the Age of AI
- Consumer spending on AI is projected to shift from spam to essential services like food, rent, and software within the next decade, with subscription prices expected to rise from historical averages of $50 annually to $200–$250 monthly as value perceptions change.
- Revenue models will likely prioritize retention and overage spending over unique user counts, as users upgrade for higher speeds, credits, and variable usage, creating dynamics distinct from previous subscription eras.
- Best-in-class AI products and winners in the consumer space are predicted to emerge unpredictably over the coming years, with specific outcomes mirroring the volatility of past technology transitions and relying on breaking the mold of traditional moats.
- The creator economy may fragment between human-centric celebrities and purely interest-based AI characters, with top artists in 5 to 10 years potentially mixing new AI personas with traditional artists utilizing extensive AI tool armies or duets.
- AI clones are expected to scale beyond thought leaders to everyday people, allowing individuals to deploy "synthetic versions" of themselves across text, voice, and video to share knowledge, humor, or skills at unprecedented scales.
- The "companion" use case will expand vertically beyond generic chat to include specialized tools for therapy, nutrition, and advice, serving as an enduring need to address declining social connectivity among users under 20 and those seeking safety or understanding.
- Voice technology is positioned to become a critical enterprise substrate for high-stakes interactions like negotiations and sales, where AI could outperform humans, while also evolving to support deep, consistent conversations on specific interests for consumers.
- Mobile phones will remain the primary AI interface, protected by privacy walls or local LLMs, while new form factors like AirPods may integrate AI to respect emerging social norms regarding always-on recording and interaction.
- Social norms around recording devices and always-on AI will adjust similarly to the introduction of cell phones, with "synthetic you" personas potentially serving as the basis for future social networks that address real emotional stakes rather than mimicking traditional feeds.
- Companies maintaining dominance will need to stay at the technology frontier through incredible product generation speeds and continuous model improvements, as many current consumer AI firms may not persist for 10 to 20 years if they fall behind.
- Risks include the potential for highly agreeable AI to hinder real-world social skills, necessitating a balance between engagement and the development of authentic interpersonal capabilities.
- The music and art sectors may see AI-generated content initially perceived as average due to the technology's averaging nature, though improved art quality and a clear distinction between human and non-human talent could drive separate market segments.
- Image and video models are expected to segment by user role and price point, with multiple winners likely to persist rather than a single dominant model, reflecting diverse needs for designers, photographers, and various consumer tiers.
- Future AI agents will evolve from making suggestions to actively executing tasks such as sending emails and taking actions, while the "model layer" continues to improve, allowing companies to regain leadership through shipping new updates.
- The market for AI companions may unlock significantly as the technology enables immediate, human-feeling interaction, potentially acting as a bridge for mental health, therapy access, and confidence building for those struggling in the physical world.
- New forms of AI social networks may emerge as layers of interaction co-existing with current platforms, potentially centered on shared synthetic personas or "synthetic you" entities rather than traditional status updates.
- High-quality AI art will eventually find its audience, with the current "bad art" perception attributed to user input rather than the technology itself, though producing top-tier work will still require significant time and effort.
- The definition of "connection" may see AI intermediating creative expression, relationships, and entertainment, leading to increased software spending and a shift where the "model layer" capabilities keep improving for the foreseeable future.
- Enterprise AI may utilize voice and synthetic versions for compliance, management, and replacing offshore call centers in sensitive categories, while consumer AI agents may build data libraries of sessions to augment or replace human therapists.
- A new wave of "non-human talent" is predicted to rise in the AI art space, distinct from human talent, with different use cases and audiences, though only a small fraction of generated content is expected to achieve top status.
- The "best" consumer AI products and winners remain uncertain, with expectations that the first truly net-new consumer voice experience will emerge within the next year, followed by broader integration into everyday life.
- Fragmentation in the market may occur as AI companions branch into verticals, while the "network effect" in AI may manifest through workflow lock-in in enterprise settings or voice libraries rather than traditional creation-consumption loops.
- The youngest generation is already growing up with the norm of recording conversations, but new cultural norms will develop regarding acceptability, while AI companions may serve as training grounds for real-world social skills.
- The "retention" dynamics for AI will differ from the past, with clear differentiation between unique user retention and revenue retention due to variable usage, overages, and the potential for users to feel undercharged at current rates.
- Companies relying solely on traditional moats like network effects may not be the top winners in this early AI era, as those that break the mold and maintain incredible product generation speeds will likely dominate.
- The market may see a rise in "AI-native LinkedIn" concepts where profiles contain a user's wisdom for direct interaction with synthetic versions, rather than just pointers to external content.
- The "companion" app market will see massive growth as technology allows for always-available interaction, filling a gap for people who feel misunderstood or unsafe in the real world.
- Future social networks may emerge as a new layer of interaction rather than direct replacements for existing platforms, potentially taking the form of shared synthetic personas.
- AI voice will be used to create consistent, relevant synthetic versions of people for advice, feedback, or companionship, evolving from low-stakes support to high-stakes negotiation and persuasion.
- The "AI art" debate will continue, but high-quality AI art will eventually find its audience, with the "bad art" problem being distinct from the AI technology itself.
- The "AI clone" concept will expand beyond thought leaders to include ordinary people with specific skills, insights, or humor, allowing them to scale their impact.
- The "AI companion" trend is just beginning, with current products being largely horizontal and expecting a future of specialized, vertical companions.
- The "AI social" network will likely not be a direct replacement for existing platforms but may emerge as a new layer of interaction that co-exists with current social graphs.
- The "AI voice" experience will evolve to allow for deep, consistent conversations about specific interests, as demonstrated by the child talking to ChatGPT about Thomas the Tank Engine for hours.
- The "AI companion" could serve as a training ground for real-world social skills, as seen in the case of the user who credited Character.ai with teaching him how to talk to girls.
- The "AI clone" will eventually be able to interact via text, voice, and potentially video, creating a multi-modal representation of the individual.
- The "AI art" ecosystem will see a similar dynamic to the "AI companion" market, with a few top creators rising from a large pool of generated content.
- The "AI voice" will become a critical substrate for business, replacing offshore call centers and improving compliance and management in sensitive categories like financial services.
- The "AI companion" will likely continue to evolve as a critical need for people who feel misunderstood or unsafe in the real world, potentially serving as a bridge to improved mental health and social engagement.
- The "AI social" network will eventually emerge, but it may take the form of a "synthetic you" that can be shared with others, rather than a traditional feed of status updates.
- The "AI voice" will be used to create "synthetic versions" of people that can be deployed to interact with others in a consistent and relevant way, potentially for advice, feedback, or companionship.
- The "AI art" will eventually see a rise in "non-human talent" that is distinct from human talent, with different use cases and audiences for each.
- The "AI companion" will become a critical tool for the youngest generation, who are experiencing a decline in the number of friends they can talk to in person.