Builders Unscripted: Ep. 2 - Ashe Magalhaes, Founder of Hearth AI

YouTube AI Channels Events

Summary

OpenAI’s Builders Unscripted Ep. 2 interviews Ashe Magalhaes on using Codex 5.4 to ship agentic tools in public and how her Slack-native workflow turns one-line prompts into products within days.

Builders Unscripted spotlights the stories behind real projects and the mindset that makes them possible: you can just build things. In this episode, Romain Huet, Head of Developer Experience,...
Original Article
View Cached Full Text

Cached at: 04/21/26, 03:34 PM

TL;DR: Ashe Magalhaes explains how she uses Codex 5.4 as a solo founder to ship agentic tools in public, why she pivoted from solar race-cars to “relationship intelligence,” and how her Slack-native workflow turns one-line prompts into full products within days. ## From Solar Cars to Satellites: The Builder Gene Ashe’s career began far from SaaS: two years in Stanford’s solar-car team—designing telemetry, CAD models, and ultimately driving the hand-built vehicle 3,000 km across the Australian outback at 90 km/h. The experience, she says, “set my bar for rigor.” Stints at Airbnb on ML infrastructure and at NASA building small satellites followed. “I treat engineering as the tool that lets an idea be born—whether it’s a satellite, a car, or a poem.” ## 2022: Prototyping the First Agentic CRM Hearth AI launched in 2022 as the first “agentic” CRM. In 2023 the team moved to GPT-3.5 while “begging friends to beta-test on extremely brittle infrastructure.” The goal: let AI reason over a user’s entire relationship graph—who you care about, why, and what next action keeps the connection alive. Early users saw bots answering inside Slack before “agent” was a mainstream word. ## Thesis: AI Should Expand Human Connection Ashe credits neurosurgeon-writer Paul Kalanithi for crystallizing her mission. “At the end, I won’t care about the satellite or the race car; I’ll care about how deeply I felt connected.” Hearth’s north star became “relationship intelligence”: AI that augments memory and intention so people feel more present with one another. ## Build-in-Public Stack, 2024 Edition Today Ashe ships live on ashe.ai, a backend disguised as a “secrets” page. Workflow in one sentence: 1. Drop idea into Codex 5.4 2. Publish MVP on aesthetic.video (embedded Gemini vector search for “every laugh in this clip”) 3. Open a dedicated Slack channel per product; 5.4 monitors logs, fixes bugs, and can split the repo into a standalone open-source site on request Stripe templates, Oura integrations, and animation libraries live in the same monorepo; she swaps models—Opus 4.6 or 5.4—without touching infra. ## Speed as a Creative Medium “Velocity is my paintbrush.” A solo founder can now iterate like a racing team: real-time telemetry (Slack alerts), pit-stop repairs (@5.4 fix), and public lap times (shipping tweets). The result, Ashe claims, is sustainable software she herself uses daily—built, observed, and refined in full view of her audience. Source: [OpenAI – Builders Unscripted Ep. 2](https://www.youtube.com/watch?v=flweA_I-VKE)

Similar Articles

Builders Unscripted: Ep. 3 - Matias Castello, Product Leader at Alchemy

YouTube AI Channels

This is an interview with Alchemy Product Leader Matias Castello, who shares how a non-engineer background transforms work through AI (Codex and GPT), including code review, product documentation, and project management, and demonstrates a workflow where AI autonomously executes development.

Codex-maxxing

Hacker News Top

Jason Liu shares how he uses OpenAI's Codex for knowledge work beyond coding, leveraging durable threads, voice input, and steering to integrate coding agents into his broader workflow.

Harness engineering: leveraging Codex in an agent-first world

OpenAI Blog

OpenAI describes an internal experiment building a production software product with zero manually-written code using Codex agents, achieving approximately 10x faster development with 1.5M lines of code written by AI in five months. The team learned that effective agent-driven development requires engineers to focus on systems design, scaffolding, and feedback loops rather than writing code directly.