Julian Lehr, The Case Against Conversational Interfaces
In Episode 003 of Hyperlegible, host Packy McCormick talks with Julian Lehr about his recent essay The Case Against Conversational Interfaces and why natural language might not be the ideal way to interact with computers. Julian explains why conversational AI should complement rather than replace graphical user interfaces, and shares insights about his unique writing process.Julian is one of the most creative writers on the internet. Go read his work at julian.digital.Hyperlegible is sponsored by our friends at Readwise. You can find links to all of our conversations, and all of the episodes we discuss, at readwise.io/hyperlegible.Episode Highlights:(00:10) - Julian summarizes his essay on conversational interfaces and why they're inefficient (02:31) - Why writing is different: "a thinking process" rather than a speed-optimized workflow (03:08) - The "Pass the Butter Test" and the ideal human-computer relationship (05:04) - The privacy vs. experience trade-off with AI interfaces (07:10) - How Julian's visual approach to writing differs from traditional writers (10:07) - Julian's unique writing process: starting with pen and paper, switching between tools (11:22) - Using Figma as an underrated writing tool for visualizing content (12:24) - How Julian incorporates AI into his writing process, using ChatGPT as a thinking partner (15:18) - Discussion of AI at the OS level and predictions about browser development (18:38) - AI as a complement rather than a replacement, inspired by StarCraft gaming interfaces (21:01) - Sam Altman's always-on AI workflow and potential applications for writers (30:58) - Why Julian took a two-year break from writing and his publishing goals for this year (32:31) - The status-signaling aspect of the "thanks to" section in essays (36:12) - Julian recommends reading his essay "The Power of Defaults" (36:55) - Recommended reading: Kevin Kwok's "The Arc of Collaboration" (38:15) - Key takeaway: Think of AI as a complement that unlocks new possibilities rather than replacing existing workflowsEssays and Podcasts Mentioned:Julian Lehr: The Case Against Conversational InterfacesJulian Lehr: The Power of DefaultsKevin Kwok: The Arc of CollaborationDavid Perell: How I Write with Sam AltmanBig thanks to Jim Portela for editing!
--------
39:07
Hyperlegible 002: Utsav Mamoria, How to live an intellectually rich life
Utsav Mamoria recently wrote How to Live an Intellectually Rich Life on his Substack, Tumse Na Ho Paayega. It blew up, breaking out of containment in India and spanning the globe to the tune of 1,100 likes at the time of recording. For good reason: Utsav combines philosophy, mathematics, biographies, personal experience, and hand-drawn sketches to create a map – quite literally – for living an intellectually rich life. He takes us on a journey through Moradoom, Igamor, and Evermore, before arriving at Luminspere, the Mountains of Knowledge. His one sentence takeaway: Consistency trumps everything.I loved reading the essay, and was pleasantly surprised to enjoy our conversation even more. You can find Utsav on X at @utsavmamoria and subscribe to Tumse Na Ho Paayega here.I asked Utsav for some recommendations. Substack to Follow: Ted Gioia’s The Honest BrokerFavorite of his own pieces other than this one: Why we understand time wrongIf you enjoy Hyperlegible, subscribe and leave us a rating. We have some great conversations scheduled and I want to bring this to as many people as we can.
--------
41:00
Hyperlegible 001: Tina He, Jevons Paradox: A personal perspective
For the first episode of Hyperlegible, I talked to my friend Tina He (@fkpxls on twitter) who writes the excellent Fakepixels, which she recently brought back to life after a four year hibernation and on which she’s dropped gems each week since.Last week, Tina wrote an essay called Jevons Paradox: A personal perspective about something surprising she’s noticed: AI is causing a lot of people to work more, not less. Since you can now do more with each hour, the opportunity cost of each hour not worked is higher! The treadmill spins faster and faster. Read it, and subscribe to Fakepixels while you’re there: https://fakepixels.substack.com/p/jevons-paradox-a-personal-perspectiveIf you're wondering how (or whether) to compete in the age of AI, Tina's personal perspective will help. Please let us know what you think and share your favorite essays with me @packym on twitter.
--------
40:20
Ryan Delk: To Fix the School System, Build Schools
In April 2022, Packy wrote a Deep Dive on Primer:https://www.notboring.co/p/primer-the-ambitious-home-for-ambitiousThree years later, a lot has changed. Primer is building schools. The opportunity is both bigger and more challenging than it was then.In this conversation, Packy and Primer CEO Ryan Delk discuss the need to fix in K-12 education. Ryan shares insights on Primer's transition from a software-focused company to a vertically integrated Microschool model. He also addresses regulatory hurdles, the market potential for innovative educational solutions, and the vision for making quality education accessible to all families.If you want to build a vertically integrated company to solve a really big, important problem, you need to listen to this.Chapters00:00The Importance of K-12 Education Reform02:18The Shift from Software to Microschools04:04The Pivot to Microschools06:28Building a New Education Model09:03Scaling and Integrating Education Solutions11:43Balancing Control and Innovation in Education14:18The Role of Software in Education17:15Future of Education and Personalization21:56The Future of Education Funding24:14AI Tutors and Their Limitations26:12The Shift in Educational Structures Post-COVID28:59Addressing Concerns About School Choice31:19Breaking the Narrative of Private Schools33:58The Vision for Free Education36:49Navigating the Incumbent Education System39:15Market Opportunities in Education41:19The Future Landscape of Education43:09Growth Projections for Primer
Italia