Signüll: Most People Are in the Stone Ages of AI | The a16z Show

The conversation delves into the accelerating pace of technological advancement, particularly in AI, and its profound impact on human culture and individual development. The speaker notes that technology cycles are becoming increasingly complex as they delve deeper into human psychology, with current advancements focusing on personality development in AI models. A primary challenge highlighted is making the power of AI models more accessible and useful to the average person. The speaker suggests that improving the Net Promoter Score (NPS) of AI hinges on making essential AI-driven services and products significantly cheaper. The discussion touches upon the speaker's unique ability to connect diverse topics, from AI and big tech to dating markets and product development, drawing parallels between life and technology through a computer science lens. This approach is described as translating internal prompts into words that resonate with others, sometimes eliciting strong reactions. The rapid acceleration of the world, likened to hitting a "100x speed" button in Sim City, is a recurring theme. Events that happened recently feel like they occurred a decade ago, driven by technology as the fuel for this rapid change. This leads to a contemplation of humanity's place amidst this technological surge: are we progressing spiritually and intellectually, or are we merely "Neanderthals with iPhones"? The speaker firmly believes that technology should facilitate intellectual, spiritual, and personal growth, enhancing self-understanding and enabling individuals to experience themselves better. This aligns with humanity's history as tool-builders, with every tool contributing to progress. The conversation then shifts to the future of AI relationships, with the speaker predicting significant, currently unimaginable, changes in the next five to ten years as AI facilitates deep human connection. A key observation is that most people are not utilizing AI beyond basic tasks, despite its advanced capabilities. The "stone age" of AI perception and usage is contrasted with the billion-plus users, who are not tapping into its full potential. The challenge, as acknowledged by OpenAI, is to make AI more accessible and useful, with agents being a step in this direction, though still primitive. The speaker draws a parallel to Shakespeare's ability to capture essence in brevity, advocating for making AI more accessible and useful. Regarding entrepreneurship in the age of large AI labs, the speaker advises focusing on personal passion and genuine interest in a problem space, rather than solely on the technology itself. Fun and enjoyment are deemed crucial for sustained effort and company building. The Bhagavad Gita quote about not being entitled to the fruits of labor emphasizes the importance of enjoying the process and the problems themselves. The discussion contrasts two archetypes of consumer founders: the highly technical builders, akin to those at OpenAI, and the "gentle builders" or "consumer philosophers" who are students of culture. The former are seen as "willing things into existence," while the latter use technology as their canvas. The complexity of developing AI personalities, a technically challenging endeavor, is contrasted with the Web 2.0 era where founders architected "delivery vehicles" for human interaction. Today, the focus is on designing the essence of human personality and intelligence, making AI the "actual thing" rather than just a conduit. Claude is highlighted as an example of an AI model that feels more "artisan" and soulful, contrasting with more robotic models. This "crafted" and "premium" feel, coupled with effective marketing and storytelling, has contributed to its adoption. Looking ahead, the speaker envisions AI evolving beyond conversational interfaces to become ambient entities woven into daily life, present in homes and workplaces without being explicitly perceived as chatbots. The evolution of interfaces, the integration of AI into operating systems, and the potential obsolescence of traditional applications are considered. The concept of ambient AI, akin to Google Now but with context and intelligence, is seen as a significant future vector. The conversation touches upon how people learn, with a humorous anecdote about learning through internet arguments. The value of learning from others, even through correction, is emphasized as a fundamental human trait. A stark contrast is drawn between AI's popularity in China and its unpopularity in the US, with AI having a lower NPS than "ICE" (Immigration and Customs Enforcement). The speaker proposes fixing this by making important things "cheaply, like soon." The diffusion of technology prices, exemplified by flat-screen TVs nearing zero cost, is contrasted with rising costs in healthcare, education, and housing. The speaker outlines how AI can make education and healthcare cheaper. For education, restoring student-administrator ratios and modestly increasing professor productivity could lead to annual cost reductions. For healthcare, addressing the 45% administrative overhead through AI can significantly lower costs, noting that healthcare companies are major consumers of AI models. The "moonshot" for the AI industry, it is suggested, should be to make these two sectors significantly cheaper within five years. The idea of subsidizing AI for these industries is raised, with the speaker agreeing that it's a possibility. The distinction between intelligence-bound problems and collective action problems is made, with housing cited as an example of the latter. A proposal for broader ownership of AI companies like OpenAI and Claude is introduced. The idea is that allowing ordinary people to own a piece of these technologies could foster a sense of ownership and a more positive view of AI, especially if this ownership is extended to future generations. This addresses the perception of wealth concentration in Silicon Valley. The discussion concludes with the speaker hinting at a new consumer product they are building with a small team, focusing on creating accessible, user-friendly consumer AI interfaces that work "out of the box." The emphasis is on "walking the walk" rather than just "talking the talk."