Episodes featuring How I AI
Browse episodes where How I AI appears as host or guest. View all episodes →
Nick Baumann from OpenAI joins How I AI to show how Codex, ChatGPT, and voice interfaces let non-engineers automate tasks like booking flights, managing expenses, and editing video.
Maddie Reese joins How I AI to show how she used AI coding tools to build hardware projects, including a receipt printer and a hacked pager, with no prior coding experience.
Claire Vo of How I AI reviews Anthropic's Opus 5, finding it timid, wordy and tiring to work with, yet the top scorer on her own blind benchmark for front-end and prototype work.
Claire Vo shows how Codex can control your browser and computer to handle repetitive tasks like QA testing, LinkedIn triage, and shopping research.
Alex Lieberman shows how he uses AI to produce authentic content 5-10x faster by mapping his workflow, codifying his voice, and using Claude as a co-pilot rather than a ghostwriter.
Alex Finn joins How I AI to show how stacking local hardware like Mac Studios and a DGX Spark can run AI models around the clock, automating coding, security, and research without cloud costs.
Claire Vo compares GPT-5.6 Soul and Fable across real product-building tasks like PRD writing, UI prototyping, coding, and voice agents, finding Soul better and 50% cheaper.
Claire Vo shows how to build custom AI harnesses using the Claude SDK, wrapping AI agents in structured code to automate specific, repeatable workflows reliably.
Alessio Fanelli shows how he manages autonomous coding agents from his phone using OpenAI Symphony and Linear, covering cloud infrastructure, task workflows, and a real Pokémon card pricing use case.
Claire Vo builds a repeatable benchmark framework to blind-test frontier AI models on PRD writing, prototyping, and agentic tasks, finding results that challenge common assumptions.
Eddie Kim, CTO of Gusto, joins How I AI to explain how his team shipped a production product in 10 weeks with five people by cutting process and leaning on AI coding tools.
How I AI and Claire Vo test GLM 5.2, an open-weight model from Z.ai, to see if it can match Claude Opus-level coding performance at a much lower cost.
Brian Grinstead from Mozilla explains how the company used agentic AI to find and patch nearly 500 security bugs in Firefox in one month.
Claire Vo explains how to build AI loops, which are autonomous scheduled workflows that let agents prompt themselves, using tools like Claude Code and Codex.
Ankur Goyal, CEO of BrainTrust, joins Claire Vo to explain how AI coding agents can solve complex infrastructure problems that engineers usually avoid because they are too costly or risky.
Claire Vo from How I AI reviews early access to Claude Fable 5, covering where it outperforms cheaper models and where it wastes money.
Nicole Ruiz shares how she built a Claude shopping assistant to find durable, quality goods for her family while avoiding cheap products in her small Brooklyn apartment.
How I AI host Claire Ho walks through using Google's Gemini Omni to build a photorealistic AI avatar and finish a full promotional video in 15 minutes.
Bryce Rattner Keithley shares how she built and launched a fitness app, Daily Hundreds, on the App Store in a few months using AI tools and no coding knowledge.
Clarvo from How I AI shares hands-on findings from testing Claude Opus 4.8, covering where it works well and where it falls short.