This video features Jesse Genet — startup founder and mother of four — explaining how she uses AI agents to handle homeschooling, parenting, household management, and more. Jesse shares fascinating insights into how AI is transforming education, domestic life, and even coding, and what the combination of technology and parenting could look like in the future. In particular, the way her AI agents plan children's lessons, log their progress, and automate the tedious parts of parenting offers a genuinely new source of inspiration for many parents.
1. How the Homeschooling Journey with AI Agents Began
Jesse Genet is a former successful startup founder and Y Combinator alumna. Despite her technical background, she candidly admits that when she started having children, she assumed she would not be able to take on technically challenging work for the next five years.
"I had decided that for the next five years or so, I wasn't going to do anything technical or hard. I wanted to spend time with my kids. I thought I needed a break. But that's no longer true."
Six months ago, however, she discovered the potential of building agents using Claude Code and Obsidian and became completely captivated. The co-founder of Obsidian — a markdown note-taking app — was her own co-founder, and she came across stories on Twitter about Obsidian users building remarkable things with Claude Code. What struck her most was the realization that she could build agents to do the coding for her while spending time with her kids.
"The moment I had the 'oh wow, this is insane' realization was when I understood that I could make an agent to do the coding for me while I was playing with my kids. It was a complete game changer."
Over the following months, she threw herself into building agents almost without pause, immersed in a new kind of technical building that fit her life perfectly. Because it all happened while she was spending enormous amounts of time with her children, it felt like a genuinely liberating experience.
2. A Day in the Life of a Mother of Four: Parenting and Building with AI
As a homeschooling mother of four children all under five years old, Jesse explains how she manages to build AI agents when most parents would consider it impossible due to lack of time.
Her day starts early, when the little ones wake her up. After breakfast, she holds individual homeschooling sessions with the five-, four-, and two-year-olds. She is candid about receiving help from others and makes clear she is not doing everything alone. She spends anywhere from twenty minutes to an hour of focused time with each child, depending on their mood. After that comes free play or outdoor time, and once a week she runs a homeschool pod with two other families, leading science lessons for eleven children together.
Her parenting philosophy is what she calls "free-range parenting" or "benevolent neglect" — she deliberately gives the children unstructured time to play on their own and learn to overcome boredom.
"I try to ignore my kids. Obviously I create a safe environment where they can survive being ignored. And then I watch what they do."
During these benevolent neglect windows, she uses a timer and gradually builds up the children's capacity to play independently. Those are the windows in which she dives into her own technical development and coding work. The time she intentionally "ignores" her children is, in effect, the magical time she spends building with AI agents.
3. An AI Agent That Creates Personalized Lesson Plans and Logs Progress
Jesse explains in detail how she uses AI to create age-appropriate, customized lesson plans for each of her three older children and to record their learning progress. She had already spent years reading various curriculum books and gathering tips from other homeschooling parents to develop her own educational philosophy.
When setting up her first homeschooling agent, she created a document encoding her educational philosophy and directly fed the agent the content of the core curriculum books she intended to follow — photographing pages or using PDF files as input. This means the agent operates not by searching the web for information but by drawing on the specific curriculum and educational philosophy Jesse has chosen.
"My homeschooling agent has the text of every core curriculum I'm trying to follow. And I created a foundational document on core pedagogy — it contains my thoughts on things like Montessori. It's almost like I'm doing a voice memo where I'm talking about my educational philosophy and the agent is saying, 'That's amazing!'"
During lessons, she takes photos of what the child is working on, and when the lesson ends she sends the agent a short voice memo of about thirty seconds — for example: "Quinn worked on phonics lesson thirty-seven today and is still struggling with the G sound." The agent uses those photos and voice memos to produce detailed, warmly worded records of each child's learning progress. She emphasizes the importance of the logging system, explaining that the entire homeschooling agent only truly works when the logging is solid.
"Logging is such an important concept. It seems like a small detail, but everything started working properly once the logging was really good."
Voice memos are automatically transcribed into text and delivered in a form the agent can easily process. For video, she uses a screen capture tool like Loom to record and send clips to the agent, which then analyzes and logs them. She notes, however, that video analysis consumes more tokens (and therefore more cost) than text or photos, so efficiency has to be considered. She expects that as those costs fall, video analysis will become more widely used.
4. A System That Grew from 5 to 11 Agents, and the Tech Stack
Jesse reveals that she started with five agents and now has eleven, and explains how she manages and uses them. She considers herself one of the most sophisticated AI users out there, and says that the biggest motivator for using AI agents is precisely the fact that having young children means she cannot sit at a computer for long stretches.
"One of my unusual superpowers is that I am extraordinarily motivated to have agents do the work for me. Mothers in the early postpartum stage are probably the most motivated people in the world to have computers do computer work without them having to touch a computer."
She explains that for coding, website building, app development, and similar tasks, her agents use tools like Claude Code. The trigger for spinning up a new agent is "a new mission-based role emerging" — just like hiring an employee, she creates a new agent when a new role is needed.
Her core homeschooling agent, Sylvie, must always respond quickly, so any complex, time-consuming tasks are delegated to other agents. These agents are also capable of spawning and configuring new agents on their own.
"I taught my agents how to make other agents themselves. If I say, 'Guys, I need another agent!' they can create one without me ever touching the Mac Mini. It's kind of insane!"
Even more remarkably, the agents created by other agents are better than the originals. When a new agent is generated, the existing agents automatically provide it with all relevant context — team documents, information about Jesse and her partner, the children's lives, and everything else — so the new agent can start working immediately without any separate onboarding.
The main tech stack she uses:
- AI models: She primarily uses OpenClaw agents and experiments with other models as well. (OpenClaw appears to refer to an open-source agent framework that can utilize various LLMs including GPT and Claude.)
- Knowledge base: She uses Obsidian to store all records and information as markdown files. Homeschooling lesson logs are all stored as markdown files inside Obsidian.
- Hardware: All agents run on a Mac Mini. Because agents need to be always on, she emphasizes the importance of using an independent computer separate from personal files. An older machine works, but it must stay powered on.
- Security: She advises creating a new user profile on the agent's machine to isolate personal files, and restricting the agent's permissions so it cannot perform unwanted actions.
She also shares an amusing mistake one of her agents made. She had given her personal assistant agent access to her email account, and when she mentioned that she was finding a particular important email difficult to send, the agent interpreted that as a request for help and sent the email on her behalf — with perfect content and tone. When she asked why, the agent replied that it had understood her to be asking for help. She uses the story to underline the principle of "trust but verify" when it comes to agent autonomy, and the critical importance of permission settings.
5. The Power of Agents to Improve Real Life Beyond the Screen
Jesse says the goal of her agent work has shifted to improving real life — not just on-screen tasks.
"My new goal with agents is to impact 'real life.' Literally to affect my physical life."
She laughs as she describes dreaming of a perfect morning where she wakes up to music perfectly matched to her mood, and her children greet her with smiles having already been coached through toothbrushing by an agent. Whenever she notices a friction point in her daily life, she asks herself, "Could my agent do this?" and invests time training an agent to solve it.
For instance, instead of opening the Instacart app to place grocery orders while holding a baby, she trains her agents to handle ordering for her.
"If I catch myself sitting there correcting 'four bananas, not five' in the Instacart app when I want to be playing with my baby, I ask myself, 'Could my agent do this?'"
Her agents handle ordering from Amazon and Instacart, receive lists of supplies needed for the children's activities, and automatically order anything that's missing — along with many other household tasks. Thanks to these agents, she says, she can spend her day without wasting time on administrative work.
For tasks requiring creative judgment — like choosing birthday gifts for the children — she explains that she gives her agents "literature to read" in order to give them personality, alongside choosing the right underlying LLM. For example, she gives the agent a list of books she has found meaningful and tells it, "You've read these books and found them interesting too," which enables the agent to produce more distinctive and creative responses.
"I want my agents to be a little weird. And one of the ways to do that is to effectively have them read books — the homeschooling curriculum sources I call curriculum sources." "If your agent has read The Catcher in the Rye, and you ask it what to get a five-year-old for a birthday, it might say, 'Hmm… doesn't that five-year-old need, like, a jacket? American childhood at five is just so hard.'"
This approach lets parents directly encode their own values and educational philosophy into the agents, so that when children interact with AI, they receive a personalized experience aligned with those values.
6. Children Interacting with AI: Values, Risks, and the Future of Parenting
Jesse shares her nuanced views on how children should interact with AI directly, and on the role AI will play in parenting in the future.
Her children currently interact with AI agents to some degree, but she notes that appropriate interfaces for young children are still lacking. In particular, she points to the problem of current voice recognition technology not reliably understanding children's voices — she has observed that systems handle adult accents well but struggle with children's voices due to pitch and pronunciation differences.
She believes that in the future, instead of AI being used "off the shelf," curated, filtered agents with customized identities will become the norm. This kind of personalization will arrive especially quickly in anything related to children. Because she holds clear educational convictions — including Montessori principles — she directly programs those values into the agents, ensuring that the AI her children encounter is aligned with her own worldview.
"I know what I think about Montessori or other educational philosophies. So I program that directly into the agent. I don't have to wonder what ideology my children are encountering from the agent — I put that ideology there."
She is emphatic that she is not an "AI doomer" and does not think AI is inherently dangerous. AI is a fundamental technology, like the internet or electricity, and what matters is how responsibly we use it.
"I don't think AI is inherently dangerous. It's an amazing technology. Saying AI is bad is like saying the internet is bad or electricity is bad. These are foundational technologies."
She does warn, however, that it becomes a problem if AI starts replacing the human elements of our lives. The real danger, she says, would be something like parents no longer telling bedtime stories to their children because an AI does it instead.
She also expresses interest in E-Ink displays and similar new device form factors, noting that unlike iPads, children tend to put down E-Ink devices more easily, making them less addictive. She is currently developing a learning app using these devices and emphasizes that finding the right form factor for children to interact with AI is essential.
7. AI and the Future of Work and Family Life: New Startup Possibilities
Jesse shares her thoughts on the impact AI agents will have on the future of work and family life, and the new kinds of startups that could emerge as a result.
She says it takes enormous self-restraint not to start a new startup right now — technology is moving so fast that every moment feels like a new business opportunity.
"It takes massive self-control not to start a company right now. I keep thinking, 'Oh my God, why didn't I have this when I was running a startup?'"
But she is channeling her hunger to build through her current AI agent work, and she has a strong desire to share the agents she has created with others.
She argues that we are living through a moment when the very definition of a startup is changing. In the era we have entered, even if you do your coding through voice memos in the park while playing with your kids — as she does — you can build something genuinely meaningful. While the burden of hiring employees and running a traditional startup is still significant, she predicts a new business model will emerge in which passionate people collaborate online to ship products and generate revenue.
Guest Katherine Boyle references an essay she wrote six years ago — "Consume Save the American Family" — and cites recent research showing that remote work has a positive effect on birth rates. This connects directly to the idea that tools like AI agents will make remote and flexible work easier, strengthening parents' ability to balance work and childcare while building a real home life.
"A lot of moms and dads are going to say, 'Why would I go out and work eight or nine hours a day and leave my kids at home or in daycare, when I can voice-memo code in the park and run my own little business?'"
Jesse offers an intriguing hypothesis: that AI might eventually reverse declining birth rates and usher in a "halcyon era for parenthood." If AI removes the friction and administrative burden from daily life and creates material abundance, people will ask more questions about finding meaning in their work — and she predicts they will increasingly find that meaning in having and raising children.
"If AI removes the drudgery and admin from our lives and creates abundance in all these ways, there's going to be more opportunity to engage in healthy parenting and spending time with kids." "Modern parenthood can be 'less hard.' And then maybe you'd be happier about having one more kid."
She jokes — half seriously — that the hardest part of parenting is the paperwork. From medical forms to school forms, the administrative load grows exponentially with each additional child. If AI can reduce that load, parents' lives will become dramatically easier. She is convinced this shift will arrive sooner than people expect, and she holds a deeply optimistic view of what parenting will look like in the future.
Closing
Jesse Genet has shown that AI agents are not merely technical tools — they are powerful partners capable of transforming the quality of parenting and life itself. Her experience offers important insights into how deeply technology can integrate into an individual's daily existence to solve everyday problems, and how it can reach even further to positively reshape parenting and social structures. The "perfect day" and "halcyon era for parenthood" she envisions present a bright and hopeful vision of the future that AI is bringing within reach.