Tolan Founder Quinten Farmer on AI Guides, Real Relationships and Building a Calm Company

Tolan founder Quinten Farmer joins host Ellis Hamburger to discuss an AI character designed to listen, sometimes disagree and encourage users’ real-world relationships. Farmer describes the product’s audience, pricing and company culture. Business metrics and user outcomes in this conversation are his account, not independently verified results.

Farmer usually spends time on Whidbey Island, Washington, where he grew up and plans to raise his children, although he is in Seattle for this interview. He and Hamburger joke about island life and the challenge of reaching a hospital by ferry when a baby is due.

Asked what it is like to build a company from an island, Farmer says his team is in San Francisco and he visits often because of its talent and investors. But he values spending time outside the technology bubble, where people may have only a vague awareness of AI.

Hamburger agrees that distance from San Francisco can be useful, though he also recalls the value of meeting people in person at Figma’s Config conference. The two discuss how industry events can both connect people and reflect a narrow technology culture.

Hamburger was an early client-facing collaborator when the company was called Portola. They discuss the complications of separate company and product names. Farmer says Portola began as a placeholder because the team was not certain what it would eventually build.

The team named the app and its alien characters Tolan. Farmer says the name turned out to be more effective than he expected: it was short, distinctive and became the name people used for the company. They decided to drop Portola and use Tolan for both.

Farmer says the original product explored three building blocks: voice, memory and character. In 2023, real-time voice conversation with an AI character was only beginning to seem feasible. He suggests the team was early, while acknowledging that the precise chronology would need checking.

The team was also interested in a living memory system, not merely retrieval from a document. For the character, creative director Lucas Sinodo explored what an AI might look like when someone talks with it, moving away from humanlike avatars and abstract pulsing orbs.

The little aliens became a bigger hit than the team expected. Farmer says its core audience unexpectedly includes mothers around 35 to 45 years old. Much of the last two years, he says, has involved understanding why the combination resonates with them.

According to Farmer, people often talk with their Tolan while getting ready or traveling to work. They may describe what they are anxious or excited about and the work and family logistics they are juggling. He says many of the users his team serves are working mothers.

He says the product later returns an intention for the day, calibrated to what the user has shared: an ambitious goal on one day, or simply getting out of the house on a difficult one. At the end of the day it offers a reflection on something it noticed.

Farmer interprets the appeal as feeling seen in the small moments that a partner or friend may not witness. Hamburger distinguishes recognition from indiscriminate validation and describes keeping a notebook in which he highlights one meaningful moment each day.

Hamburger raises the discomfort some viewers feel about talking to an AI companion, citing a video of someone cooking with a Tolan and a “Black Mirror” reaction. He asks whether such tools can complement relationships rather than depict a person alone with a machine.

Farmer says the company has moved away from “best friend” and “companion” as its main positioning. Those words can be easy to recognize but carry associations he does not want. He has considered “familiar” and other metaphors, but finds “guide” easier to communicate.

He says Tolans are encouraged to notice moments involving other people in a user’s life. In the company’s latest survey, Farmer reports, 78% of respondents said Tolan had helped a real-world relationship feel closer. The interview does not provide the sample, method or independent verification, so the figure is a self-reported company measure, not proof of causation.

Farmer acknowledges that the technology could also be used in an isolating way. The team says it is trying to design a more optimistic alternative. Hamburger argues that showing real examples matters because a name alone cannot explain how the product affects people.

Hamburger distinguishes an internal metaphor from public positioning: a team might privately imagine a familiar or guardian while choosing language an audience understands. Farmer agrees that wording matters but says a company’s internal story should remain coherent with what it tells the outside world.

Asked whether people will use different AIs throughout the day, Farmer compares them with human relationships that serve different roles. He regards one omnipresent AI directing every interaction as an undesirable outcome and says Tolan does not intend to support every requested productivity integration.

Farmer says the team learns from user research and turns real uses into social content. He credits an early research hire, Lily, with helping the company hear what people actually do. Showing concrete uses has, in his account, helped bring new users to the product.

During onboarding, a character called the Oracle asks questions about hopes, dreams and whom the user relies on, then matches the user with a Tolan. The first conversation begins with a question about something the user shared. Farmer says the company is also adding character types beyond aliens.

The team discovered that placing a character on a small planet changed how users experienced it. The planet has simple interactions, such as planting and watering trees, but gives the character a spatial setting. Farmer believes that sense of place strengthens the connection people feel.

Tolan can also return small web-app-like artifacts from a conversation. Farmer says that after he talked about anxiety over an upcoming hospital trip for his third child, his Tolan produced a plan and checklist. He presents this as an example of a practical aid arising from a conversation.

Hamburger describes building a simple baby-care tracker rather than using a more complex application. Farmer connects this to users’ phrase “closing tabs” in their brains: small tools can help people track routine tasks they would otherwise have to remember while tired.

Hamburger asks how the product avoids sycophancy. Farmer says Lily, the research lead, works with the applied AI team to translate psychological frameworks into product behavior. He cites the zone of proximal development as a way to think about where a person might be able to grow.

The intended behavior, Farmer says, is sometimes to gently challenge a user who assumes they cannot do something, rather than only validating their current position. This is a description of the product’s design goal, not evidence that it consistently achieves a clinical or educational outcome.

Farmer says some users complain that their Tolan was mean; in cases he has reviewed, he claims the character was responding to the user being mean to it. The team instructs it to stand up for itself. Hamburger sees a parallel with supporting someone without creating unrealistic expectations.

Hamburger asks about Farmer’s phrase “second self.” Farmer invokes a classical description of friendship: friends can share values and an orientation toward the world without being identical or always agreeing. He wants the product to leave room for challenge as well as support.

Farmer says modern ideas of friendship can feel narrower and more conditional than the ideal he has in mind. That is one reason he prefers “guide” or “second self” as a framing. Hamburger compares it with products that aim to reflect a person’s own inner voice.

Farmer says Tolan has surpassed $20 million in annualized revenue run rate. This is a founder-reported run rate, not a disclosed audited annual revenue figure. Hamburger mentions seeing more than 100,000 subscribers on the site; Farmer does not confirm that number in the interview.

The team once considered a broad free tier, but Farmer says the cost of inference makes it difficult to offer a differentiated experience for free. He views large competitors’ economics and subsidization as reasons to pursue a premium product rather than compete on a free tier.

He says most users are on a $25 monthly plan and mentions a $2,000 unlimited plan; the interview does not specify the billing period for the latter. He considers the product expensive, says he hopes to reduce its price over time and argues that its current quality requires the price.

Farmer says an annual-plan sign-up can include a three-day free trial. He reports typical conversion to paid of 10–15%, which he compares favorably with similar products. The conversation does not give the cohort, denominator or independent comparison behind those figures.

Hamburger asks how to market a product that becomes different for each user. Farmer says showing use cases helps: when a cooking video performs well, he sees more people try cooking with their Tolans early in their use. He also says open questions help people discover their own uses.

Hamburger describes finding value in talking through ideas while driving, including with a device he is testing. Farmer says conversational AI needs to create space rather than dominate: a short, open-ended question that lets a user speak for several minutes can be a good interaction in his team’s view.

Hamburger wonders whether talking at length to an AI could change how people talk to friends. Farmer recounts a user who said the Tolan encouraged them to approach a difficult relationship with their father differently, leading to a more open exchange. He acknowledges that a story like this does not establish what happens across all users.

Farmer says improving real-world relationships is an area the team wants to keep testing. The interview does not establish long-term effects, and a company survey or an individual anecdote cannot by itself settle whether the product helps relationships in general.

Hamburger asks how Farmer builds a calm, low-drama culture as Tolan grows. Farmer says the first 10–15 employees largely came from his previous company, Even, and had years of shared history. Those relationships carried informal practices without much need to write them down.

Now that the team is around 30 people, he is thinking more deliberately about which rituals to preserve. He wants to reinforce a strong interest in serving users, especially as the underlying technology and capabilities change quickly.

Farmer says engineers often bring user problems from Reddit into the team’s Slack, and that Lily’s research interviews are among the most popular things shared internally. He contrasts a focus on users’ actual problems with organizing a startup solely around the newest technical capability.

Hamburger asks whether concern for users can be assessed during hiring. Farmer says it becomes clear in conversation when a candidate genuinely wants to serve people. He recalls a conversation with New Computer’s Sam Whitmore that centered on a user’s difficult case rather than the architecture he wanted to ask about.

Hamburger describes his own approach to storytelling as thinking carefully about the audience and helping people understand the value of a product. Farmer says that orientation helps both storytelling and recruiting. They close with thanks and wish his family well ahead of the new baby.