A box of 24 doughnuts never reached Jai Bhatia’s office, but the follow-up email did: a young programmer explained his failed delivery attempt and offered the founder a trial of his software.
Bhatia, the Indian-origin co-founder and chief technology officer of Turnstone, shared the message publicly last week, calling it “one of the coolest emails I’ve ever got”. The post drew attention to the sender’s attempt to make an unsolicited business approach personal.
The email came from Logan Hogg, who described himself in the accompanying handwritten letter as a 20-year-old self-taught programmer. He said he had been in San Francisco, had tried to find the office and was returning to London the following day.
The request at the end was specific: try the software for a week. It was a product pitch, rather than a confirmed job application or an announcement that Hogg had been hired.
What did the Jai Bhatia email say?
The message explained that Hogg had visited with doughnuts and a handwritten letter but had been unable to find the office, so he attached the letter instead.
The original post shows both the short email and a photograph of the handwritten note. Hogg’s letter expressed admiration for Bhatia’s work and explained why he wanted to make contact.
The email then shifted from the attempted gift to a commercial proposal. Hogg offered to show what he had built and suggested that Bhatia’s team could try it free for a week. He claimed the product could reduce spending on AI tokens and time spent debugging.
The handwritten note also proposed lunch or dinner while the sender was in San Francisco. That was another request, not a record of a meeting. The public message does not supply an acceptance, so neither a proposed meal nor a product demonstration should be presented as a completed encounter.
Those claims were part of the pitch. Bhatia’s praise for the email does not establish that the software delivered the claimed savings or that Turnstone became a customer.
The screenshot’s email date and the date of the public post are also separate. The message is dated 7 November; Bhatia shared it publicly later. That distinction avoids presenting the office visit as something that happened on the day an account of the post appeared.
Who is Jai Bhatia?
Bhatia is listed as Turnstone’s co-founder and chief technology officer in Y Combinator’s company directory.
The directory identifies Aryan Vij as co-founder and chief executive. It places the company in San Francisco and lists it in Y Combinator’s Winter 2026 batch.
Bhatia’s founder profile lists computer science at the University of California, Berkeley, while Vij lists a Berkeley engineering background. Their company profile describes work on AI tools and interfaces that allow users to control software through conversation.
That background explains the connection between the recipient and a pitch about AI development costs. It does not establish a prior relationship between Hogg and Bhatia. The email was presented as an unsolicited approach.

Was the message asking for a job?
The request visible in the email was to demonstrate a product and arrange a trial, not to announce an employment agreement.
That distinction matters because stories about unusual approaches to founders are often interpreted as recruitment stories. Here, the sender proposed a possible customer relationship: a team trying software that he believed could solve a problem for it.
Bhatia’s public reaction concerned the email. The post did not announce a job offer, an investment or a signed customer contract. There is no basis in the message itself for treating any of those outcomes as completed.
What is a cold email in this case?
It is an unsolicited message to a potential business contact. Hogg used it to explain a missed in-person delivery and propose a software trial. The expression describes the approach, not an endorsement of the product or a successful sale.
The physical delivery attempt and digital follow-up were parts of the same approach. The doughnuts supplied the anecdote; the trial request supplied the commercial purpose.
What made the pitch specific to its recipient?
It combined a personal explanation for the contact with a defined request directed at an AI company’s technical team.
The sender did not simply ask Bhatia to respond to an unspecified opportunity. He explained the office visit, attached the letter and proposed a limited period in which the team could assess his software.
The product claim was also connected to the recipient’s field: AI token costs and debugging. Whether the claim was accurate would require testing, but the problem described was narrower than a promise to improve every part of a business.
These are observable features of the email, rather than proof of a universally successful formula. Another recipient could respond differently, and an admired message can still lead to no commercial agreement.

What is the difference between attention and a verified business outcome?
Attention means the recipient noticed and publicly reacted to the approach. A business outcome would require evidence of an agreed trial, purchase, investment or employment decision.
The distinction is particularly relevant when a social post circulates beyond the original exchange. A positive caption is evidence of the caption’s sentiment; it is not a substitute for a customer announcement or a statement that a product has been tested.
The same applies to the sender’s description of himself and his software. The letter provides his account of his age, training and motivation. The email provides his claims about what the tool could do. Those claims should retain their attribution rather than become independent technical findings.
For founders reading the story as a sales example, the most concrete lesson is the structure of the request: the recipient was offered something specific to evaluate. That observation does not require turning the attempted gift into a recommendation to visit offices uninvited or spend money on elaborate outreach.
What does the pitch’s claim about AI tokens actually mean?
It concerns the units an AI model processes, rather than a demonstrated saving that can be read from the email itself.
Google’s token-counting documentation describes tokens as units of text used to process input and distinguishes a token count from a word count. Its documentation also separates an input-token estimate from the actual input, thinking and output tokens processed.
That distinction matters to a cost claim. A shorter input is one measurement; the complete usage and the price attached to that usage are further questions. A team evaluating a proposal would need a defined comparison before it could establish what changed.
In this case, the email did not publish a test method, a baseline bill or results from Turnstone’s workloads. It offered a trial in which a team could investigate the product. The claim remains the sender’s proposal rather than an independently measured finding.
Where does the exchange fit within advice on startup outreach?
It is an example of a business approach, while broader training treats outreach as a process that needs more than an attention-grabbing opening.
Y Combinator’s September 2024 Startup School lesson, presented by group partner Aaron Epstein, addresses cold emails for attracting customers. Its published outline covers the outreach funnel, finding contacts, responses and examples.
Those topics identify different stages. An email can reach someone, receive a reply and lead to a demonstration without yet producing a sale. The distinction is useful when interpreting Bhatia’s reaction: it confirms a positive response to the message, while later commercial stages remain unconfirmed.
The lesson predates this exchange and is not a review of Hogg’s product or an endorsement of his visit. It is a separate primary resource for readers interested in the mechanics of outreach beyond one widely shared anecdote.
What can young founders take from the exchange?
They can distinguish a clear proposal from the attention it receives and keep any claims about the result tied to evidence.
A proposed one-week trial creates a question that a team can answer: whether it wants to test the tool. It does not establish the eventual price, security conditions, data access or purchasing decision. Those matters would need their own agreement if the conversation progressed.
Y Combinator’s startup library offers primary material on company-building and sales for people seeking a broader introduction. Bhatia’s post remains one individual exchange, rather than a measured study of which approaches produce customers.
Other business reporting and diaspora coverage place the founder’s story in a wider context.
What happens next?
Any development beyond the public reaction would need confirmation from the people or company involved. The post does not supply the result of a product trial or a subsequent deal.
For now, the verified outcome is narrower: an unsuccessful attempt to deliver a letter and doughnuts became an email that its recipient chose to share approvingly. The proposed software trial remains a proposal in the material made public.
