She vibe coded an iPhone app and launched it to the App Store with zero coding knowledge
Bryce Ratner Keithley, a non-technical talent and recruiting professional, built and launched a fitness app called Daily Hundreds to the Apple App Store using AI vibe coding tools including Replit, Claude, and Lovable — with zero coding knowledge. The app features AI-generated anthropomorphic animal exercise videos created by combining Gemini-generated images with real workout footage processed through Higgsfield's Cling model. Her journey illustrates how beginner's mindset and AI tools are enabling non-technical people to build production-ready software.
Summary
Bryce Ratner Keithley, whose career has been entirely in talent and recruiting, built a consumer fitness app called Daily Hundreds and successfully launched it to the Apple App Store without any software engineering background. The app originated from a pandemic-era habit of doing 100 reps of a single exercise per day, and Bryce wanted a tool that would suggest a different exercise each day and allow her to log her reps.
Bryce began building in October using both Lovable and Replit simultaneously, ultimately sticking with Replit because a friend worked there as a director of engineering. She described her process as leveraging a 'beginner's mindset' — not knowing the boundaries of what was possible actually helped her push further. Key workflow discoveries included using Replit's plan mode before making changes, which prevented the AI from going off in unintended directions. She also relied heavily on screenshots and very literal, precise prompting to guide the AI.
One of the most distinctive features of the app is its library of AI-generated anthropomorphic animal exercise demonstration videos. Bryce developed a multi-step pipeline: she creates animal characters in Gemini with very precise positional prompts (leveraging her background as a barre instructor for accurate physical queuing), films herself performing the exercises, and then combines the animal image with her workout video using Higgsfield's Cling 3.0 motion control model. This process requires significant iteration, as the AI frequently misinterprets positioning instructions.
For getting the app into the App Store, Bryce initially heard from technical friends that she would eventually need a developer to unblock her. However, by February, new AI models made self-service possible. She used a layered Claude workflow: plain Claude as a 'technical architect' to create a plan and provide context, Claude Code as the 'software engineer' to write actual code, and the terminal to execute commands directly. She spent approximately 25-30 hours in a single weekend and got the app approved on her second submission after fixing three issues: a mislabeled parental control checkbox, a non-functioning Sign in with Apple (required for iPad compatibility), and a missing account deletion button.
In the broader conversation, Bryce reflected on talent and hiring implications, arguing that people who remain territorial about their traditional roles or resist seeing their function through a new lens will struggle. She emphasized that for technical roles especially, the human value-add has shifted — finding a working solution fastest is no longer the differentiator since AI can do that. She recommended the books 'What Got You Here Won't Get You There,' 'How Women Rise,' and Daniel Pink's 'A Whole New Mind' as relevant reading for navigating this career moment.
Key Insights
- Bryce argues that not knowing the boundaries of what's technically possible is actually an advantage — she acquired Railway as her infrastructure without knowing what it does, illustrating how non-technical users are now discovering and purchasing developer tools that were previously sold exclusively to engineers.
- Bryce found that Replit's plan mode was a critical unlock — without it, asking the AI to make UI changes directly would cause unpredictable, hard-to-reverse results, but planning first allowed her to collaborate more intentionally with the AI before any code was written.
- Bryce's barre instructor background — specifically her ability to precisely cue physical positioning — turned out to be directly transferable to prompting AI image generation models, allowing her to describe animal starting positions with the exactness needed to get consistent, usable results in Higgsfield.
- Bryce used a three-layer AI workflow to get her app into the App Store: plain Claude as a high-level technical architect providing the plan, Claude Code as the hands-on software engineer writing code, and the terminal for direct execution — with Claude also reviewing Claude Code's output at each step.
- Bryce contends that technical candidates who focus only on finding a working solution fastest in interviews are missing the point, because AI can already do that faster — and that engineers who don't recognize their role has fundamentally shifted to encompass the full suite of tools will face growing irrelevance.
Topics
Transcript
[0:00] I built an app called Daily Hundreds. I opened Lovable and Replet actually on the same day and left the simplest prompt. It was incredible to me that I could tell these AI tools, I want this. And it spit out a very basic minimum viable product of it. I asked you the other day, I was like, "Are you in test flight?" And you were like, "Yeah, I was in Test Flight." And so now it's in the app store. You got it approved. It's ready to go. We have anthropomorphic animal demos. This is my favorite part. and we're going to see some pretty amazing ones. But can you walk us [0:31] through how you generated this?…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
GPT-5.6's video editing via Codex is genuinely one of my favorite new workflows
A product manager describes using GPT-5.6 with Codex to automate video editing for social media clips. By simply dragging a file and providing natural language instructions, the AI generated five polished, fast-paced hype videos from a long conference talk recording, dramatically reducing the time-intensive manual clipping process.
Theoretically Intelligent vs. Practically Effective: Why GPT-5.6 Sol Beats Fable for Product Work
An executive contrasts two AI models (Fable and Soul/GPT-5.6 Sol), arguing that Soul is superior for product work because it prioritizes practical effectiveness over theoretical intelligence. The speaker values the ability to ship products to customers and understand end-user goals over theoretical sophistication.
Build a harness when the same workflow needs the same setup and the same outcomes, every time
Building a harness for repetitive workflows allows you to be more prescriptive about job execution, resulting in greater efficiency, consistency, and better outcomes. Rather than explaining requirements to an AI agent each time, a harness lets you use a simpler interface like pasting a link while the agent already understands the intended task.
GPT 5.6-Sol vs. Claude Fable: Why OpenAI’s new model crushes my benchmark
Claire Vo compares OpenAI's new GPT 5.6 models (Soul, Terra, Luna) against Claude's Fable using her custom "How I AI" benchmark, finding that GPT 5.6 Soul excels at practical product work, prototyping, and natural communication, while Fable is theoretically intelligent but pedantic and difficult to collaborate with.
Context offloading is an underrated AI use case
The speaker highlights context offloading as an underrated AI use case, where AI serves as a safety net for routine cognitive tasks like email management and personal finances. Rather than adding new capabilities, AI reduces anxiety about missing important information or making mistakes by handling monitoring tasks, thereby freeing up mental bandwidth.