Y Combinator
A startup accelerator that funds and advises early-stage companies.
Y Combinator shows up as a source of blunt advice, network and fast fundraising. Pingo's strong metrics booked about 100 investor meetings and closed its round in 3 days, with partners urging it to push UGC. Replika pivoted after Paul Graham called its app a demo. Lesson: use the accelerator to focus on the one lever that works, and look for niches like behavioral health that model companies avoid.
- Koby's Rupa background
- Usage numbers
- Mostly failed to raise
- Where the market is going
- Product-market fit, positioned in AI
- Revenue rising, not usage
- Letters of intent from pilots
Raised about $6.5M total; Flybridge led seed
From Sunflower founder Koby Conrad
What was said about Y Combinator
1What Sunflower took from its time at YC
Build AI where model companies lack outcome data and avoid the liability. Koby Conrad heard in a YC talk that OpenAI staff stay out of behavioral health because of legal liability and unclear targets. A startup that tracks outcomes like relapse, days sober and loneliness can optimize the whole system around them without training a new model.
The opening is where the agent is only part of the solution, so tooling and function calls can be tuned to those outcomes. OpenAI and Anthropic don't know whether users are less depressed or staying sober. Quote: "they don't know what to optimize for."
Pitch investors on where the market is going, backed by revenue growth. Sunflower's first deck leaned on Koby's Rupa background and usage numbers, and it mostly failed. The deck that raised money after YC described where the market is going, showed product-market fit, positioned Sunflower in AI, and showed revenue rising rather than usage.
It also included letters of intent from pilots. Sunflower has raised about $6.5 million in total, with Flybridge leading the seed. Quote: "you can't raise funds based off of a playbook that came out of 2025."
2Pingo's experience with YC
Strong metrics plus the YC network made fundraising fast. Pingo's founder found YC worth it for the network, the blunt advice and easy fundraising. Strong metrics booked about 100 investor meetings, and the round closed in 3 days.
YC partners told Pingo to push the one lever that works. The advice from partners was to double down on UGC, the channel already working for the app.
3Replika and the Paul Graham pivot
Paul Graham called Replika's app a demo for something bigger, prompting a pivot. Replika's team went through YC with a restaurant recommendation app. Paul Graham said it was a demo for something more exciting, which led them to pivot to a messenger with bots.
Lessons from 1,000+ YC startups: Resilience, tar pit ideas, pivoting, more | Dalton Caldwell (YC)
Dalton is a managing director and group partner who has worked across 21 batches. Its Request for Startups, about 20 ideas including ERP software, exists to change founders' idea diet and not to limit what YC funds.
Paul Graham called Replika's app a demo for something bigger, prompting a pivot. Replika's team went through YC with a restaurant recommendation app. Paul Graham said it was a demo for something more exciting, which led them to pivot to a messenger with bots.
Lessons from working with 600+ YC startups | Gustaf Alströmer (Y Combinator, Airbnb)
Gustaf has been a group partner for almost six years and has worked with over 600 startups. YC has funded over 130 climate companies, cannot pick outlier winners, and focuses on preventing known failures through weekly office hours and group sessions.
Strong metrics plus the YC network made fundraising fast. Pingo's founder found YC worth it for the network, the blunt advice and easy fundraising. Strong metrics booked about 100 investor meetings, and the round closed in 3 days.
YC partners told Pingo to push the one lever that works. The advice from partners was to double down on UGC, the channel already working for the app.
Sam Altman on Building OpenAI & Betting on the Impossible
Altman says its philosophy of iterative deployment, technical people in charge and betting on young ambitious founders made OpenAI possible. He compares it to the band that sold few albums but influenced every musician, and says he spent about 10 years watching thousands of its companies.
Build AI where model companies lack outcome data and avoid the liability. Koby Conrad heard in a YC talk that OpenAI staff stay out of behavioral health because of legal liability and unclear targets. A startup that tracks outcomes like relapse, days sober and loneliness can optimize the whole system around them without training a new model.
The opening is where the agent is only part of the solution, so tooling and function calls can be tuned to those outcomes. OpenAI and Anthropic don't know whether users are less depressed or staying sober. Quote: "they don't know what to optimize for."
Pitch investors on where the market is going, backed by revenue growth. Sunflower's first deck leaned on Koby's Rupa background and usage numbers, and it mostly failed. The deck that raised money after YC described where the market is going, showed product-market fit, positioned Sunflower in AI, and showed revenue rising rather than usage.
It also included letters of intent from pilots. Sunflower has raised about $6.5 million in total, with Flybridge leading the seed. Quote: "you can't raise funds based off of a playbook that came out of 2025."
How Wabi is Building “YouTube for Software” with a Platform for Creating Mini Apps
Replika's team went through it with a restaurant recommendation app. Paul Graham called that app a demo for something more exciting, which led to a pivot to a messenger with bots.
How I built a $500k/mo ai app (so you can just copy me)
Worth it for Pingo because of the network, blunt advice and easy fundraising. Strong metrics booked about 100 investor meetings and the round closed in 3 days, and partners advised pushing the one lever that works, UGC.
Tony Xu of DoorDash: Surviving 1,000 Days of Startup Hell
DoorDash spent the summer of 2013 at YC testing three numbers: whether consumers would pay $6, restaurants would accept 15%, and drivers could earn a workable wage. Demo day and fundraising were not the goal.
I speedran an ai app from 0 to $1m/yr in 10 months
Invested in Sunflower, and OpenAI staff told Koby in YC that they avoid behavioral health over legal liability and unclear targets.
Brian Armstrong: When Washington Tried to Kill Coinbase
Its advice led Armstrong to call signups and add a buy button, and its network got Coinbase a bank account after many banks hung up on him. He went through the program solo after a hasty co-founder split and raised a $600K seed round at demo day.
Behind the founder: Drew Houston (Dropbox)
Houston got in through a Hacker News demo video that caught Paul Graham's attention, then had to find a co-founder fast. Money and the product came quickly after Demo Day.
Brian Chesky’s new playbook
Michael Seibel of Y Combinator advised Brian without being an advisor, investor or board member, asking only that Brian help other founders in return.
An inside look at Deel’s unprecedented growth | Meltem Kuran Berkowitz (Head of Growth)
Deel has a partnership with Y Combinator that gives qualified members access to it. Closed communities like this are less crowded because access is the hard part.
More in Investors & studios
Investors & studiosSequoia Capital
A venture capital firm that backs early-stage technology founders.
8 sources · 18 tipsBlue Throne
An app acquirer that buys working consumer apps for seven and eight figures and scales them.
4 sources · 6 tips
Investors & studiosBending Spoons
An Italian company that acquires digital businesses and runs them on one shared platform of proprietary tools.
2 sources · 19 tipsSocial
A bootstrapped app studio that buys apps after their viral peak and grows them with AI ads and testing.
1 source · 10 tips
Investors & studiosTeledyne
A conglomerate founded in 1960 that bought niche businesses, ran an insurance investment arm, and repurchased its own stock.
1 source · 17 tipsVolo Ventures
A venture firm investing in consumer apps and AI-powered consumer products.
1 source · 9 tips
Investors & studiosFirst Round Capital
A seed-stage venture capital firm that backs early technology startups.
4 sources · 5 tipsBenchmark
A venture capital firm that invests in consumer and marketplace startups.
3 sources · 4 tips