EO StudioI Built a General AI Agent After School at Age 13 | Flowe AI, Michael Goldstein
CHAPTERS
- 0:00 – 0:30
School bans AI, but students use it anyway—spark for building Flowe
Michael explains how his school technically doesn’t allow AI, yet AI tools have become a daily reality for students. That tension—and the rapid rise of products like ChatGPT—helped inspire him to create Flowe.
- •Schools may discourage AI use, but students still rely on it for notes and answers
- •The rise of mainstream AI tools opened a “new bridge” for him to build on
- •Michael introduces himself as the founder of Flowe AI
- 0:30 – 1:00
What Flowe AI is: a “general AI agent” that completes real tasks
He describes Flowe as a broad, action-oriented agent that can execute practical tasks like booking flights or canceling subscriptions. He frames it as a step toward more capable, AGI-like systems.
- •Flowe aims to “get things done,” not just chat
- •Example tasks: booking flights, canceling subscriptions, making presentations
- •Positioned as a glimpse of the future and closer to AGI
- •Michael notes he’s 13 and based in Toronto
- 1:00 – 1:30
From idea to prototype: building without adult developers and using AI tools
Michael shares that when he first tried to recruit developers, most said no—likely due to his age. He decided to build it himself using AI tools and started weekly coding lessons to become fully independent.
- •Initial outreach to developers failed; age likely affected responses
- •He began building Flowe on his own with help from AI tools
- •Weekly coding lessons to strengthen fundamentals and reduce reliance on others
- •Early-stage reality: still learning while shipping
- 1:30 – 2:00
“Build, build, build”: balancing a kid’s life with entrepreneurial drive
He reflects on advice to live a “normal kid life,” but emphasizes his desire to keep building. He also describes a friend group with similar entrepreneurial energy.
- •He prioritizes building over a conventional schedule
- •Has a close friend group aligned around entrepreneurship
- •Mindset: consistent creation and iteration
- 2:00 – 2:31
Early maker projects and support system: weather balloons, family, and friends
Michael describes launching weather balloons that reached 100,000 feet and even got CNN coverage during the solar eclipse. He credits his friends and family for understanding his builder mindset and being excited about Flowe.
- •Weather balloon launches as a hands-on engineering project
- •First launch received CNN attention during the solar eclipse
- •Friends weren’t surprised by his startup ambitions
- •Family was excited and supportive when he launched Flowe
- 2:31 – 3:01
First customers: emailing users, fixing bugs, and the first Stripe sale
He recounts extracting user emails, sending a discount, and then quickly needing to fix a broken code issue. The moment he received his first Stripe sale felt like a breakthrough he immediately shared with family.
- •Pulled user emails from Supabase and emailed a discount offer
- •Immediate customer feedback revealed a glitch that he fixed
- •First Stripe sale felt surreal and validating
- •Celebrated with friends and parents right away
- 3:01 – 3:31
Going viral and being 13: attention, opportunities, and credibility challenges
Michael explains that his age helped his launch video perform well and opened doors to interviews and events. At the same time, it can reduce trust and makes recruiting especially difficult without capital.
- •Age likely boosted curiosity and virality of the launch content
- •Opportunities: interviews, visibility, and recognition
- •Credibility is harder—some people don’t trust a young founder
- •Recruiting is difficult with limited capital
- 3:31 – 4:32
Momentum from exposure: Founders Inc., AI summit invites, and new collaborators
He shares notable outcomes from the attention: acceptance to Founders Inc. as the youngest participant and an invitation to a top AI summit in Paris. He also explains that key team members eventually reached out to him first.
- •Accepted to Founders Inc. (youngest accepted)
- •Invited to a major AI summit in Paris
- •After months, recruited two developers and a strategy helper
- •Team members initiated contact, reversing earlier recruiting struggles
- 4:32 – 5:02
How kids actually use AI at school: mostly ChatGPT and little awareness beyond it
Michael observes that AI interest is widespread among students, but tool usage is narrow. Most classmates rely almost entirely on ChatGPT and don’t explore other options unless they’re deeply involved in AI.
- •Students broadly use AI daily, not just his friend group
- •ChatGPT dominates; other tools are less known
- •Discovering broader AI tooling requires deeper involvement
- •School culture: AI is common even if not fully embraced institutionally
- 5:02 – 5:32
The core product gap: browser automation to eliminate repetitive human work
He explains that his deeper understanding of AI revealed an unmet need: automating tasks in the browser, like forms and workflows. He positions Flowe as filling this gap that major chat models don’t directly solve end-to-end.
- •He identified a “big gap” in automated task execution
- •Focus on browser automation and form-filling as a high-impact use case
- •Belief that current major tools don’t fully deliver this workflow automation
- •Efficiency for everyday life (including parents’ time) as a key motivation
- 5:32 – 6:33
Roadmap and goals: fix bugs, add features, and aim for $10K MRR
Michael outlines near-term improvements—more features, smoother performance, and fewer bugs. He also shares a public revenue goal of $10,000 in monthly recurring revenue, acknowledging it may take time.
- •Plans to improve reliability and resolve many existing bugs
- •Wants a richer feature set within the next year
- •Set a public goal of $10K monthly recurring revenue
- •Understands growth may take years but wants meaningful progress within a year
- 6:33 – 7:05
Long-term outlook: Flowe’s role depends on AI progress and a desire to give back
He connects Flowe’s future to how AI capabilities evolve, aiming for the product to become part of everyday life by automating what can be automated. He closes with a longer-term motivation to help people and give back.
- •Product direction tied to the pace and shape of AI evolution
- •Vision: become a daily-life tool that automates routine tasks
- •Sees Flowe as part of a broader future of capable agents
- •Personal goal: help others and contribute positively over the long run
