AI & I

Brandon Gell, Every’s COO, Kieran Klaassen, who leads Cora, and Austin Tedesco, Every’s head of growth, unpack a live doc editor built for people and AI to write together. They talk agent-native design, real-time co-writing, planning docs, multi-agent editing chaos, creative workflows, and why some AI-written docs are made more for other AIs than for human readers.

Dan Friedman, operator who builds service+software companies like Moxie, and Sam Gerstenzang, incubator partner who grows real-world+software businesses to multimillion revenue. They discuss building med spas and funeral services, their slow incubator that scales then hands off leadership, choosing unsexy defensible verticals, and how they use AI agents for customer discovery while rejecting synthetic call validation.

Alex Mathew, a 17-year-old Alpha High senior and founder building Berry, an AI-powered plush companion for teen mental health. He describes life at an AI-driven school with no traditional teachers. Short takes on anti-cheating, Pomodoro blocks, Gen Z attitudes toward AI and social media, and how teens build projects, startups, and alternative paths to college.

Thibault Sottiaux, head of Codex at OpenAI, leads product and research; Andrew Ambrosino builds Codex app features and automations. They discuss Codex’s rapid growth, a new ultra-fast model and desktop app, why a GUI beats terminals for visibility and multimodal workflows, and how speed, infra, and automations reshape coding and make code review the next bottleneck.

Darin Fisher, veteran browser engineer focused on UX and reliability, and Ben Goodger, head of engineering known for browser design, discuss Atlas, OpenAI’s agentic browser. They talk about everyday tasks AI can take over, how built-in agents change browsing, UI choices for discoverable AI features, and the engineering challenges of building a new browser with AI coding tools.

Natalia Quintero, Head of AI consulting at Every who built an AI project manager called Claudie and has guided 100+ organizations, walks through designing and iterating an agent that automates onboarding, status updates, and project workflows. She covers mapping high-leverage tasks for private equity, linking proprietary context to speed memo drafting, and the rituals that unlocked rapid experimentation and adoption.

Andrew Wilkinson, a serial entrepreneur and cofounder of Tiny, shares his innovative use of AI tools like Opus 4.5 and Claude. He discusses his transition from early coding attempts on Replit to now feeling empowered by AI, likening it to having a vast team of engineers at his disposal. Wilkinson reveals his creation of 'Deep Personality', an AI relationship counselor, and talks about automating personal and work tasks, including a custom email client and daily outfit recommendations. He reflects on how these advancements are reshaping his approach to software acquisitions.

Nir Zicherman, entrepreneur and CEO of Oboe, dives into the challenges of self-directed learning with AI. He highlights how traditional LLMs might not understand when learners are struggling or losing interest. Nir discusses the nuances between passive and active learning, arguing for the need for purpose-built platforms like Oboe that generate personalized courses. They also explore how different subjects require varying teaching methods and technologies, and the importance of keeping learners engaged without overwhelming them. Nir shares his own experience using Oboe to fill learning gaps.

Felix Rieseberg, a product engineer at Anthropic, shares insights about Claude Cowork, a user-friendly AI tool designed for everyone. He explains the innovative design that separates 'Tasks' from 'Chats' and the async queue system that enhances productivity. The discussion covers practical demos like email drafting, calendar audits, and competitive analysis. Felix also highlights the importance of 'Skills' for customization, and the principles behind agent-native architecture, hinting at a future where knowledge workers can seamlessly interact with AI.

Reid Hoffman, co-founder of LinkedIn and an early investor in OpenAI, shares his bold predictions for AI by 2026. He discusses how agents will expand beyond coding into diverse areas, reshaping enterprise work. Hoffman warns of increasing negative sentiment around AI as it faces backlash for societal changes. He emphasizes the potential of AI in biotechnology, forecasting significant breakthroughs. The conversation also touches on the evolution of programming skills, stressing orchestration as a new frontier for developers.