Issue No. 37Week ending Sunday, September 13, 2026434 episodes · 1825 articles
The Throughline ↓
The Podcast Summary.

40 hours of podcasts, in 5 minutes.

Host

Claire Vo

Host, How I AI

Claire Vo appears in 17 full episodes we cover on How I AI. Host of How I AI. Below is what each conversation covered, with a key takeaway per article. Every quote in the articles is verbatim and timestamped to the source video.

17 episodescovered
77 articleswith timestamped quotes
AIHow I AI

Stripe built a company brain: Meet Kai

Sharadh Krishnamurthy, engineering manager at Stripe, joins Claire Vo to break down Kai, Stripe's internal AI agent used by over 10,000 employees weekly. He explains why Stripe decided to build rather than buy their com…

  • Stripe deployed its internal AI agent, Kai, to more than 10,000 active employees every week. Read →
  • Over 10,000 Stripe employees use Kai every week, creating more than 2,000 custom skills by converting ad-hoc chat sessions into saved workflows. Read →
  • Over 10,000 Stripe employees use Kai, the company's internal AI agent, every week to run operational tasks and inspect internal data. Read →
AIHow I AI

GPT-6 Astra blew away every one of my benchmarks

Claire Vo shares her early access hands-on review of OpenAI's GPT-6 Astra model, demonstrating how its computer use and coding capabilities exceed prior models like 5.6 Sol and Fable. She walks through production workfl…

  • Claire Vo ran OpenAI's GPT-6 Astra model inside Chrome for 1 hour and 45 minutes to autonomously test a staging preview branch. Read →
  • Claire Vo spent six months attempting to build an automated product intelligence system across models like Fable and 5.6 Sol without success. Read →
  • GPT-6 Astra creates functional 3D Blender assets in one shot, overcoming previous frontier model failures that output broken meshes and distorted geometries. Read →
AIHow I AI

7 Grok Bot agents I use every day

Claire Vo explains why she migrated roughly 30 active agents from OpenClaw to Grok Bot, breaking down the platform's core primitives: bots, multi-account plugins, virtual machine computer use, and routines. She details…

  • Claire Vo built two specialized Grok Bots, named 'Look Good To Me' and 'Lockdown', to automate engineering management and SOC 2 compliance. Read →
  • Claire Vo migrated roughly 30 active agents from a self-hosted Mac Mini setup running OpenClaw over to Grok Bot. Read →
  • Claire Vo migrated roughly 30 active agents from a local Mac Mini running OpenClaw to Grok Bot because maintenance overhead crushed daily usability. Read →
AIHow I AI

I built a Claude Cowork system that does a week of PM work in a day

Daniel Blum, a product manager at Melio, details his comprehensive Claude and Cowork personal productivity system that compresses a week of PM overhead into a single day. He demonstrates his automated Sunday weekly prep…

  • Daniel Blum, product manager at Melio, runs an automated morning brief using Claude and Cowork that synthesizes Granola transcripts, Slack messages, and emails into one-line summa… Read →
  • The main bottleneck stopping AI from handling product management overhead is local computer execution; agents must run in persistent cloud runtimes like OpenClaw or Hermes when yo… Read →
  • Daniel Blum runs 70 to 80 percent of his computer time through Cowork and Claude, shrinking five days of product management administrative work into one. Read →
AIHow I AI

How I manage 15 AI agents 24/7 as a solo founder | Ryan Carson

Solo founder Ryan Carson joins host Claire Vo to discuss how he manages 10 to 15 autonomous Devin agents in the cloud to ship dozens of daily PRs for Untangle. They explore automated QA workflows, the Watchdog customer…

  • Ryan Carson eliminated resumes, phone screens, and live culture calls to hire Untangle's first engineer, replacing them entirely with unedited desktop screen recordings. Read →
  • Ryan Carson ships up to 40 pull requests per day for Untangle by orchestrating 10 to 15 Devin agents running concurrently in the cloud. Read →
  • Ryan Carson runs 10 to 15 autonomous Devin agents in the cloud for Untangle, shipping dozens of pull requests every day. Read →
AIHow I AI

How I use ChatGPT to run my fashion business

Claire Vo interviews Yana Welander about her AI-native fashion startup, Yana Banana. Welander details how she leverages AI throughout the entire business process, from initial design sketches and sophisticated prompt en…

  • AI image models like Image 2.0 are exceptional at following precise visual directions from a sketch, but often render designs two-dimensional, losing realistic fabric flow. Read →
  • Yana Welander, a non-technical founder, built Yana Banana's entire e-commerce infrastructure using AI tools like Codex and Browser Use, sidestepping the need for human engineers. Read →
  • AI agents aren't killing SaaS; they're revitalizing it by interacting with existing software tools. Claire Vo admitted a "trough of despair" about SaaS, only to find agents excel… Read →
AIHow I AI

How this OpenAI engineer uses Codex + ChatGPT Work to automate everything

Claire Vo, joined by OpenAI's Nick Bowman, explores advanced use cases of Codex and ChatGPT Work, including voice-activated task delegation, multi-threaded AI assistants that manage expenses and travel, building and dep…

  • The initial "AGI moment" for most users, from 2022-2023, was through reactive chatbots that only responded to direct commands. Read →
  • AI can now interpret high-level intent like "find the best takes" and "UGC style video" to autonomously assemble polished content, moving beyond simple cut instructions. Read →
  • AI deploys live: OpenAI’s chatgpt.site feature lets you prompt ChatGPT to build and deploy functioning websites directly from the app. This isn't just code generation; it’s puttin… Read →
AIHow I AI

I hate Opus 5. It’s the best model, anyway.

Claire Vo opens up about her exhaustion with the rapid release of new AI models, proposing a future shift in focus from raw intelligence to factors like speed, cost, and open source. She then delves into a unique review…

  • In a unique "psychological comparison," Claire Vo found Claude Opus 5 views itself as a "shallow thinker" that openly depends on humans for empathy and judgment, even suggesting h… Read →
  • Claire Vo reveals her frustration with Anthropic's verbose "Claude slop," contrasting it with GPT's directness and detailing the hidden costs for builders. Read →
  • Claire Vo, feeling the fatigue of constant new AI model releases, advocates a shift in focus from raw intelligence to practical metrics: speed, cost, and open-source availability.… Read →
AIHow I AI

Local AI models explained: How to run a fleet of Mac Studios and GPUs at home

In this episode, Claire Vo and Alex Finn explore the practicalities of running local AI models at home, discussing the necessary hardware, setup processes, and diverse use cases. Finn shares his personal 'fleet' of Mac…

  • Your First Employee: Always-On AI. Alex Finn runs a "fleet" of local AI models on Mac Studios, DGX Spark, and Nvidia GPUs that work 24/7, performing high-volume, continuous tasks… Read →
  • Dependability trumps 'wow' factor: Alex Finn chose Hermes over Open Claw not for superior features, but because Open Claw frequently broke, demanding "half an hour fixing it" week… Read →
  • Alex Finn runs an autonomous 'software factory' using local AI models, primarily Claude Code, on a personal 'fleet' of Mac Studios, a DGX Spark, and Nvidia GPUs. Read →
AIHow I AI

How I run autonomous coding agents from my phone with OpenAI Symphony + Linear

Claire Vo and Alessio Finelli discuss how Alessio uses OpenAI's Symphony and Linear to manage autonomous coding agents, streamlining software development tasks and tracking costs. They also explore AI's broader impact o…

  • Track token usage ruthlessly: Don't just prompt and hope for the best. Alessio Finelli, a builder working with OpenAI's Symphony, constantly tracks token usage to pinpoint where a… Read →
  • AI is unlocking a new wave of small business creation by efficiently bridging the gap between digital tools and physical world realities. Read →
  • Stop manually prompting AI agents for every code change. Alessio Finelli argues for a shift from being an "agent prompter" to an "agent manager," overseeing a workflow that autono… Read →
AIHow I AI

Loop engineering for beginners

This episode explores the concept of AI agent loops, contrasting them with traditional prompting and explaining how agents can be designed to prompt themselves autonomously. Claire Vo details various types of loops, out…

  • Claire Vo reveals how AI agent loops automate prompting, letting agents work autonomously. Ditch manual prompts, but watch out for token costs. Read →
  • AI agent loops enable autonomous, self-prompting systems, moving beyond one-off interactions to recurring, goal-driven execution. Read →
AIHow I AI

How this startup uses AI agents to eliminate bugs and optimize infrastructure

Claire Vo and Anker Goyel discuss how AI agents are transforming software engineering, particularly for complex infrastructure and database optimization. Goyel highlights the importance of rigorous evals as a modern app…

  • AI agents are solving deep technical challenges like database query optimization by exhaustively trying different column store formats and execution engines, a task impractical fo… Read →
  • AI agents eliminate excuses for product backlogs. Anker Goyel states, "there's just no excuse to not have rigor... we don't really have a backlog," meaning every "paper cut" or pe… Read →
  • Machine learning fundamentally changes programming from defining how to build something, to defining what the successful outcome looks like. Read →
AIHow I AI

She built a Claude shopping assistant to stop buying cheap junk

Nicole Ruiz shares how she leverages AI, specifically Claude, to create a personalized shopping assistant that curates high-quality, long-lasting products, avoiding cheap junk and knockoffs. She also demonstrates using…

  • AI for Grunt Work: Nicole Ruiz uses AI, specifically Claude integrated with her email, to automate the tedious process of product returns, from finding receipts to drafting emails. Read →
  • Learn how one founder uses Claude AI to scrutinize brands for hidden problems, from PE acquisitions to bad management reviews, saving time and avoiding junk. Read →
  • Nicole Ruiz was fed up with the "crappy plastic version" of baby gear and other household items she bought online, realizing she spent too much time on maintenance. Read →
AIHow I AI

Spec-driven development: the AI engineering workflow at Notion | Ryan Nystrom

Claire Vo and Ryan Nystrom, an engineering manager at Notion, discuss how AI has revolutionized their engineering workflows. They cover an AI agent that automates meeting preparation, a system for AI-powered code genera…

  • Your engineering team's comfort making frequent changes directly links to CI speed; slow CI creates “monster changes” no one wants to review. Read →
  • Automate Low-Value Prep: Notion's "Hot Potato" AI agent autonomously gathers updates from Slack, tasks, and pull requests, generating a meeting pre-read daily at 9:00 a.m. Read →
  • Notion’s internal “Boxy” system integrates an AI agent called Codeex to automate code generation and bug fixes directly from Notion tasks. Read →
AIHow I AI

The internal AI tool that's transforming how Stripe designs products | Owen Williams

Claire Vo interviews Owen Williams, a design manager at Stripe, about "Protodash," an internal AI-powered prototyping tool he built. They discuss how Protodash solves the problem of creating realistic, data-rich dashboa…

  • Stripe's internal tool, Protodash, directly addresses the inefficiencies of traditional design reviews, moving from static presentations to interactive prototypes. Read →
  • Stripe's internal AI tool, Protodash, directly confronts the perennial bottleneck of designers being overloaded, enabling product managers to self-prototype realistic, data-rich d… Read →
  • Traditional tools like Figma hit a wall when prototyping complex data dashboards, struggling with variations like zero data states, abundant data, or internationalized text. Owen… Read →
AIHow I AI

How Sendbird became an AI-first company | John Kim (co-founder and CEO)

In this episode, Claire Vo interviews John Kim, co-founder and CEO of Sendbird, about their journey to becoming an AI-first company. They discuss Sendbird's innovative internal platforms like the 'Automator' for AI ques…

  • Sendbird's internal 'Automator' platform and 'AI skills marketplace' lets employees propose 'AI quests' and learn new AI tools. Read →
  • Sendbird built an internal AI skills marketplace, akin to an app store, where employees create and share reusable 'plugins' and individual 'skills' to solve problems. Read →
  • Forget generic notetaking; John Kim, CEO of Sendbird, uses an open-source AI assistant, 'The Gardener,' to actively enrich his personal notes stored in tools like Obsidian. Read →
AIHow I AI

How a non-technical founder built a $100K ARR meme company | Jason Levin (Memelord CEO)

In this unhinged episode, Jason Levin, founder and CEO of Memelord, shares his journey as a non-technical founder building a meme company to $100K ARR using no-code and AI tools. He and Claire Vo discuss the future of A…

  • Jason Levin, CEO of Memelord, reversed his long-held belief: AI can be genuinely funny, driving his meme company to $100K ARR using no-code and AI. Read →
  • The modern attention economy rewards entertainment above all else; your brand's content must directly compete on this metric, a thesis that drove Memelord's founder Jason Levin. Read →
  • Founders should focus on building hyper-personalized, often hardware-based, AI solutions for their own specific problems, not just scalable products. Read →
The Sunday Email

Get next Sunday's issue in your inbox.

40 hours of podcasts, distilled into one 5-minute read. Free, every Sunday morning.

One email a week. Unsubscribe with one click.