AIHow I AI
Blair Metcalf, principal UX engineer at Reddit, discusses the architecture and impact of Reddit Proto, an internal vibe-coding harness used across Reddit's product and UX organizations. He details how the team abstracte…
- Blair Metcalf built an internal tool called Reddit Proto that converts visual UI annotations directly into targeted Cursor prompts. Read →
- Executives routinely kill design prototypes because static lorem ipsum feeds and mock accounts do not match their personal interests. Read →
- Reddit built Reddit Proto, a dedicated desktop application that hides Git, terminal commands, and file directories so designers and product managers can code in Cursor without tec… Read →
AIHow I AI
Claire Vo reviews Anthropic's personal AI agent, Claude Duet, assessing its design polish, onboarding, and core primitives like artifacts and channels against competitors like Grok Bot and Muse. While praising its UI el…
- Anthropic built Claude Duet with polished UI, connectors, and artifact channels, but the agent fails to initiate contact after setup. Read →
- Anthropic built Claude Duet with clean UI polish and channel structures, but its mobile-only design locks it out of normal desktop work hours. Read →
- Anthropic structured Claude Duet around three concrete interface primitives: a dual-track task list, cross-platform artifacts, and Slack-style channels. Read →
AIHow I AI
Tech veteran and creator Brooke Lacey demonstrates how she uses Claude Code alongside terminal tools like FFmpeg and Whisper to automate short-form video editing locally in seconds without traditional editing software.…
- Brooke Lacey avoids ad-hoc prompt corrections by using a lane operating system that isolates individual product environments from global operating instructions. Read →
- Brooke Lacey replaced conventional video editors like CapCut and Adobe Premiere with a terminal-driven script running Claude Code, FFmpeg, and Whisper. Read →
- TikTok blocks direct programmatic scraping even through its official API, stopping AI agents from pulling audience feedback directly from the backend. Read →
AIHow I AI
Alex Cornell, design director on Meta's Muse, breaks down the user experience choices behind designing a consumer-first AI agent. He discusses why consumer agents should resemble messaging platforms rather than develope…
- Meta designed its AI agent Muse by abandoning enterprise developer patterns like files, diffs, branches, and code threads in favor of familiar messaging conventions. Read →
- Alex Cornell, design director on Meta's Muse, argues that building consumer agents requires unlearning deterministic UI design where engineers and designers control every pixel. Read →
- Pure generative UI creates disorientation; Alex Cornell argues consumer agents require rigid structural primitives rather than unpredictable layouts that redraw every session. Read →
AIHow I AI
Claire Vo runs Meta's new frontier AI models, Muse Wonder and Wonderful, through the How I AI benchmark across coding, agentic workflows, prototyping, and creative tasks. She finds the model exceptionally capable at age…
- The How I AI benchmark evaluates frontier models on roughly 20 practical workflows divided into five buckets: prototypes, agentic tasks, coding, writing, and creative output. Read →
- Meta's Muse Wonder swept all four agentic evaluations in the How I AI benchmark, outperforming rival models in back-end coding, inbox triage, and computer use. Read →
- Claire Vo benchmarked Meta Muse Wonder against Claude Opus 5.5 and Astra, finding its front-end design quality trailing competitors by 12 to 18 months. Read →
AIHow I AI
OpenAI Product Lead Kath Korevec joins host Claire Vo at OpenAI DevDay to showcase the practical and creative applications of ChatGPT Sites and Codex. They explore how teams use live connectors to build incident command…
- ChatGPT Sites includes integrated Cloudflare D1 SQL databases, Cloudflare R2 object storage buckets, and multi-user co-editing out of the box. Read →
- OpenAI introduced connected plugins inside ChatGPT Sites at DevDay, letting pages adapt live based on the logged-in user's identity. Read →
- Recommendation algorithms degrade when multiple people share an account, polluting curated queues with kids' media or irrelevant clicks. Read →
AIHow I AI
Claire Vo breaks down TypeSafe AI's decision model Jev, explaining how its low cost and typed output primitives (choice, score, and boolean likelihood) enable high-speed data classification. She shares real-world produc…
- Claude Code and Codex store complete user turns and session history directly on your local hard drive. Read →
- Analyzing 1,700 pull requests across ChatPRD took two minutes and cost 9 cents using fast decision models instead of heavy vector embeddings. Read →
- Claire Vo built a real-time multimodal app in one afternoon by combining OpenAI's Realtime Voice API, Jev, and the API Ninjas quote endpoint. Read →
AILenny's Podcast
OpenAI product leaders Tara Seshan and Nan Yu join Claire Vo at the Lenny and Friends Summit to discuss the realities of building AI products in a fast-moving landscape. They explore why shipping imperfect features is e…
- Working with AI research teams requires a different toolkit than standard engineering, focusing on user failure modes instead of sprint tickets. Read →
- Tara Seshan recommends that AI product teams plan only two to three months ahead instead of writing multi-year roadmaps. Read →
- Nan Yu argues that multi-agent systems fail when they exceed human cognitive limits, leading users to build a single "chief of staff" meta-agent to manage the sprawl. Read →
AIHow I AI
Claire Vo recaps the major product announcements from OpenAI DevDay 2026 in San Francisco. She shares hands-on testing results for OpenAI's new autonomous agent called Dots, the collaborative workspace ChatGPT Spaces, a…
- OpenAI introduced ChatGPT Spaces at DevDay 2026 as a shared environment for documents, slides, and HTML artifacts where humans and software agents work together. Read →
- OpenAI introduced three execution tiers for its developer models: normal mode, fast mode at 2x speed, and ULTRAFAST mode at 8x speed. Read →
- OpenAI introduced the Decisions API at DevDay 2026, creating a direct competitor to Jev for fast, constrained classification. Read →
AIHow I AI
Claire Vo and John Lindquist discuss Jev, an ultra-fast, low-cost decision and classification model from Typesafe AI. They explore architectural patterns and real-world product demos, including real-time voice task mana…
- Most voice apps stall because they wait for silence detection or an enter key before sending transcripts to a slow LLM. Read →
- Typesafe AI's decision model, Jev, executed moves 10x faster and 4x cheaper than a low-reasoning LLM in live blitz chess benchmarks. Read →
- Generative models process unstructured text to produce unstructured text, introducing latency and hallucination risks when applications only need deterministic routing. Read →
AIHow I AI
Claire Vo reviews Anthropic's Claude Opus 5.5 after stepping away from Claude models due to verbose and preachy responses. She evaluates its pricing, tone changes, safety boundaries, multi-turn agentic performance, UI p…
- Claire Vo tested Claude Opus 5.5 on four complex agentic workloads (inbox triage, backend coding, deep research, and computer use) running between 25 and 82 steps from a single pr… Read →
- Claire Vo abandoned earlier Claude models not over intelligence benchmarks, but because of excessive conversational verbosity and preachy disclaimers. Read →
- Claude Opus 5.5 writes clean, valid SVG vector code directly in text, eliminating the malformed coordinates common in older models. Read →
AILenny's Podcast
In her keynote at the Lenny and Friends Summit, product executive Claire Vo argues that the traditional feature roadmap is obsolete in the era of AI-driven coding agents. With engineering capacity no longer scarce, team…
- Claire Vo, a product executive with more than two decades of product leadership, argued at the Lenny and Friends Summit that the traditional product roadmap is dead. Read →
- AI coding agents break the traditional roadmap because writing code is no longer the scarce resource in software development. Read →
- AI coding agents make writing code cheap, but customer trust remains scarce and easy to destroy. Read →
AIHow I AI
Claire Vo runs a live, blind benchmarking session comparing OpenAI's GPT-6 Sol and GPT-6 Astra against Anthropic's Claude Opus 5.5 across real-world workflows including PRDs, email triage, frontend coding, SVGs, and vid…
- Standard AI benchmarks track spaceship games and Python scripts, but Claire Vo uses the "Barbie Bench" (generating a functional 3D Barbie fashion designer game) to test spatial re… Read →
- Public leaderboards measure synthetic trivia, while production work depends on ergonomics, speed, and day-to-day readability. Read →
- Anthropic fixed the chatty verbosity of Opus 5, but Opus 5.5 stops narrating its actions, which makes execution feel artificially slow. Read →
AIHow I AI
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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 →