This transcript captures a series of introductions and discussions from the "Solve It With Code" Discord community, led by Jeremy Howard of fast.ai. The participants, a diverse international group of software engineers, data scientists, researchers, and professionals from various fields, share their motivations for joining the course. A central theme is a widespread dissatisfaction with current AI-assisted coding practices, often described as "vibe-coding," which participants feel can lead to skill atrophy, mental laziness, and a superficial understanding of their work. They are drawn to the "Solve It" method as a more structured, human-centric approach that uses AI as a collaborative partner for deep learning and problem-solving. The transcript includes detailed testimonials from previous attendees who praise the methodology for improving code quality, fostering a deeper understanding of problems, and enhancing their overall work experience. The conversation also touches on community building, the challenges of AI hype, and the importance of a thoughtful, empowering approach to AI in contrast to what some see as the "dehumanizing" trends in the tech industry.
Joining the 'Solve It With Code' Community: Motivations, Methods, and Testimonials
10/4/25, 7:31 PM – The Learning Philosophy: New members express excitement not just about coding, but about the course's philosophy of learning and building through small, interactive dialogues with AI as a "cointelligence."
10/4/25, 9:35 PM – Key Testimonial 1 (Dani Helinko): Jeremy Howard shares a detailed testimonial from a first-cohort participant.
Takeaway: The SolveIt method positively impacts the feeling of work, deepens the understanding of the problem domain, and improves the quality of the final product.
Example: The user built a production-ready annotation app in two weeks, which evolved into a more robust, general-purpose tool because the method encouraged a deeper exploration of each component.
10/4/25, 9:40 PM – Key Testimonial 2 (Pol Avec): A second testimonial highlights the course's focus on meta-learning.
Takeaway: It's described as a course on "learning how to learn" in the age of AI, applicable even to non-programmers. It actively counters the trend of "vibe-coding" (delegating work to AI without deep engagement), which is likened to the psychologically manipulative patterns of a slot machine and can cause skills to atrophy.
Core Principle: The approach prioritizes the user's psychological health, learning abilities, and happiness, which is contrasted with the typical focus of venture-backed tech companies.
10/4/25, 11:58 PM – Addressing AI Hype: A user notes that converting friends to the course is harder now that the initial AI hype has subsided.
Key Takeaway: The community recognizes the need to move beyond talk and demonstrate the method's power by building and sharing impressive projects. The leadership acknowledges this and plans for more project-focused courses.
10/5/25, 4:33 PM – Community Building in Action: Members of an independent fast.ai study group introduce themselves. Jeremy Howard encourages them to migrate to the official fast.ai Discord server to consolidate the community, prevent fragmentation, and connect them with more learners. He then shares an invite link for everyone.
10/6/25, 11:06 AM – A Superior Development Style: A developer with a background in Clojure compares Jeremy Howard's notebook-driven style to Lisp's highly effective REPL-driven development.
Key Insight: This iterative style is seen as "way superior" to typical test-driven or manual feedback loops for refining a problem. The user is interested in bridging the gap between notebooks and production software.
10/8/25, 8:14 PM – Counteracting Mental Laziness: A common motivation for joining is a desire to combat the negative side effects of using AI coding tools.
Takeaway: Multiple users report that over-reliance on tools like Claude Code has made them "mentally lazy" or caused them to feel they aren't learning or understanding as deeply as when they wrote code themselves. They seek a better balance.
10/16/25, 6:23 PM – Resisting Dehumanization: A user expresses alignment with the idea that large tech companies are pushing AI in a "dehumanizing" way.
Takeaway: Another member adds that corporate leadership is fostering job insecurity by suggesting software roles will be irrelevant by 2026, which ironically discourages the very learning needed to adapt. The SolveIt approach is seen as a human-centric alternative.
10/16/25, 12:30 PM – Diverse and Global Community: The introductions reveal a remarkably diverse group of participants from dozens of countries (USA, UK, Australia, Germany, Benin, Brazil, Japan, Norway, etc.) and a wide array of professions, including physicists, veterinarians, archaeologists, film VFX artists, PhD students, and entrepreneurs, all united by an interest in this problem-solving methodology.