Skip to content

Podcast

What’s Next for Consumer AI? | Josh Elman Joins A16z

The a16z Show

Source ↗ ← All highlights
  • Siri Found My Dinner From A Text
    • Josh Elman recounts using Apple’s new assistant to navigate to a dinner by finding a text message with the address.
    • The assistant pulled data from messages and calendar to avoid a wrong turn a few blocks away. (Time 0:06:54)
  • Consumer AI Should Augment Everyday Life
    • AI can shift from productivity/work replacement to enhancing everyday life and personal experiences.
    • Josh Elman argues natural language interfaces and personalization let people create, customize, and explore things previously hard or impossible for consumers. (Time 0:08:20)
  • Chat Is A Launchpad Not A Replacement
    • Chat-style interfaces are a powerful starting substrate but won’t replace rich apps and experiences.
    • Elman expects chat to get users into apps faster, then users will spend time in immersive, visual experiences like games. (Time 0:13:58)
  • AI Discovery Will Use Creators And Agent Referrals
    • Discoverability for AI-native products will be multi-channel: creators, virality, search, and ‘generative optimization’ by agents.
    • Elman predicts assistants that refer out to vertical apps will boost discovery and create ecosystems. (Time 0:21:24)
  • Prioritize Retention Before Scale
    • Focus on retention over raw acquisition; get a small group to move their lives onto your product before scaling.
    • Elman says build a product people can’t go back from, then repeat and expand from that core. (Time 0:24:30)
  • Robinhood Grew With Share Giveaways And Word Of Mouth
    • Josh Elman explains Robinhood’s early growth: product-first word of mouth, strong design, Reddit buzz, then a referral that gave actual shares.
    • The giveaway-share mechanic acted like a lottery and amplified organic excitement efficiently. (Time 0:27:37)
  • Musical.ly Became TikTok Through A Creation Loop
    • Elman recounts Musical.ly’s loop: easy creation, curated feed, and cross-posting to Instagram that created influencer stars.
    • ByteDance acquired and scaled it with paid ads, but retention from personalization made those ads effective. (Time 0:31:55)
  • Design Pricing Around Inference Costs
    • Charge consumers when inference costs are meaningful; users are willing to pay for valuable AI features.
    • Elman notes easier in-app payments and advises moving cheaper inference to device while sending complex work to the cloud. (Time 0:36:45)
  • Labs Won’t Win Every Consumer Use Case
    • Large incumbents and major assistants will dominate some layers, but many consumer opportunities remain for startups.
    • Elman argues users will expect assistant-like behavior everywhere, creating openings for specialized products and experiences. (Time 0:38:51)
  • Build AI That Helps People Spend Time Well
    • Build consumer AI that helps people spend time well across life tasks like travel, health, finance, and learning.
    • Elman prioritizes consumer experiences over enterprise workflows, highlighting underexplored personal use cases. (Time 0:49:10)