Why Your AI Learning Projects Keep Fizzling Out (55 min)
- Release date: 2026-01-14
- Listen on Spotify: Open episode
- Episode description:
LLMs have made it absurdly easy to go deep on almost any topic. So why haven’t we all used ChatGPT to earn college degrees we wished we had majored in or pursued a niche interest, like learning how to name the trees in our neighborhood? I know I’m not the only one to feel guilty for well-intentioned attempts at autodidactism that inevitably peter out.Entrepreneur Nir Zicherman has a reason for this disconnect: LLMs can answer most of your questions, but they won’t notice when you’re lost or pull you back in when your motivation starts to fade.As the CEO and cofounder of Oboe, a platform that generates personalized courses about everything from the history of snowboarding to JavaScript fundamentals using AI, Zicherman has thought deeply about why the ability to access information does not automatically lead to understanding a concept. In this episode of AI & I, he talks to Dan Shipper about everything he’s learned about learning with LLMs.They get into Zicherman’s counterintuitive belief that learning is a more passive process than you’d think, the biggest blocker for most people who want to learn something new, and where AI agents currently fall short in providing a meaningful learning experience.If you found this episode interesting, please like, subscribe, comment, and share!Want even more?Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.To hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribeFollow him on X: https://twitter.com/danshipperTimestamps:00:00:00 - Start00:00:36 - Introduction00:01:49 - Why you need a dedicated AI learning app00:04:32 - The process of learning is more passive than you might think00:10:21 - Live demo of Oboe to create a course about philosopher Ludwig Wittgenstein00:16:52 - Learning works best when it comes in many formats00:28:21 - Where AI agents currently fall short in the learning experience00:34:10 - The importance of making learning feel accessible00:35:56 - How Zicherman uses Oboe to learn quantum physics00:40:54 - How embeddings spaces remind Dan of quantum mechanicsLinks to resources mentioned in the episode:Nir Zicherman: @NirZichermanLearn something new with Oboe: https://oboe.com/
Summary
- 🚀 Oboe’s Mission: Oboe is a purpose-built AI app creating on-demand, multimodal courses for any topic, making intimidating subjects feel achievable through objective-driven paths.
- 🧠 Passive & Multimodal Learning: Most learning happens passively via varied formats like podcasts and visuals, not constant chat; Oboe emulates this with quizzes and scaffolding for better retention.
- 📱 Live Demo Highlights: Instant generation of a Wittgenstein course shows progressive loading, podcasts, and adaptive formats, pulling sources while prioritizing speed and quick starts.
- 📈 User Insights & Growth: Two-thirds of users seek goal-oriented courses; post-launch updates like unlimited free creations and Series A fuel roadmap for refinements and re-engagement.
- 🔬 Quantum-AI Analogies: Discussions link quantum weirdness (Stern-Gerlach, double-slit) to LLM embeddings, questioning determinism, knowledge, and human limits in understanding AI ‘intuitions’.
Insights
Why isn’t a general-purpose LLM like ChatGPT sufficient for effective learning?
Time: 1:57 – 4:19
Category: AI Tutors & Personalized LearningAnswer: LLMs excel as information compressors for quick answers but lack the multimodal formats, passive consumption paths, and persistent scaffolding needed for sustained, objective-driven learning. A dedicated platform like Oboe builds these in, mimicking real teachers by guiding users back to goals without constant user prompting. (Start at 1:57)
Is most human learning passive and multimodal rather than intensely active?
Time: 2:35 – 7:23
Category: AI Tutors & Personalized LearningAnswer: People primarily learn through passive consumption of varied formats like reading, videos, and podcasts, piecing together rabbit holes across platforms, rather than constant active engagement. LLMs force excessive user direction, unlike teachers who structure content proactively. (Start at 2:35)
How does objective-driven motivation transform intimidating learning goals into achievable paths?
Time: 5:19 – 6:12
Category: AI in Skill DevelopmentAnswer: Over two-thirds of Oboe users input high-intent goals like ‘master mortgage basics’ or ‘pass a test,’ revealing a common gap where people know their end objective but lack steps. AI platforms succeed by mapping personalized, milestone-based journeys that build confidence and momentum. (Start at 5:19)
Why do LLMs struggle with long-term learning despite long contexts?
Time: 9:23 – 32:52
Category: AI Tutors & Personalized LearningAnswer: Even advanced LLMs lose focus, context, and motivation adaptation over extended sessions, failing to reassess like a teacher who reads the room or re-engages lapsed students. Dedicated platforms need agentic guardrails for persistent, adaptive paths without user micromanagement. (Start at 9:23)
What makes AI-generated courses feel lightweight yet substantial for real progress?
Time: 13:58 – 23:26
Category: AI Tutors & Personalized LearningAnswer: Oboe generates fast-loading, progressive courses with intros, podcasts, quizzes, visuals, and embedded formats tailored to content type, avoiding overwhelm while providing quick wins and scaffolding. This balances autonomy for deviations with always returning to the core objective. (Start at 13:58)
Can AI platforms bridge the re-engagement gap in fragmented learning routines?
Time: 29:01 – 35:33
Category: AI in Skill DevelopmentAnswer: Users drop courses due to life interruptions and context loss; future AI needs notifications, cross-device support, and smart recaps to reignite interest without guilt. Oboe’s unlimited free generations enable iterative refinement, turning ephemerality into flexible longevity. (Start at 29:01)
Do quantum experiments reveal parallels to the ‘black box’ magic of LLM embedding spaces?
Time: 41:20 – 46:25
Category: AI & Human IdentityAnswer: Just as quantum measurements collapse probabilistic states unpredictably (e.g., Stern-Gerlach, double-slit), high-dimensional embeddings capture implicit human patterns invisible to us, enabling predictions we can’t explain. This challenges Newtonian views of knowledge, broadening it to include unarticulated human intuitions. (Start at 41:20)