Oct 21, 2025 · 38m · y-combinator
Startup Advice: AI GTM, Pivoting & How To Hire · Y Combinator
Gustaf Alströmer · 11m spoken Nicolas Dessaigne · 9m spoken Brad Flora · 9m spoken Pete Koomen · 3m spoken Video Submitter · 18s spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this Y Combinator Office Hours episode, general partners break down essential strategies for early-stage founders, focusing on AI go-to-market execution, market targeting, pivot decisions, and early hiring thresholds. They emphasize hands-on founder execution, continuous customer iteration, and maintaining operational leanness until reaching genuine product-market fit.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The partners as informed peer 6.1 Guest teaching 2.9 Guest disagreement 1.4 The partners pushing back 1.9
speaking balance: gold is the partners, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 26:30 Rejecting technical difficulty as pivot rationale
Nicolas strongly rejects the premise of running away from technically difficult ideas, arguing that high barriers create the strongest startup moats.
Hardest push from the partners ▶ 24:23 Challenging the premise of good ideasBrad challenges the conventional concept of 'good startup ideas', asserting that anything short of great is inherently a bad bet for a venture scale business.
Biggest teaching moment ▶ 8:50 Caveating mid-market learning adviceNicolas educates the panel on Algolia's learnings, explaining that certain problems exist purely at the enterprise tier, making down-market pivots ineffective.
The partners hold their own ▶ 28:03 Brad sharing tactical modular engineering strategyBrad demonstrates hands-on founder mastery by recounting how Perfect Audience circumvented an overwhelming tech hurdle via API front-ends and custom billing.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The partners as informed peer | Guest teaching | Guest disagreement | The partners pushing back | Why |
|---|---|---|---|---|---|---|
| Go-To-Market Strategies for AI in Legacy Industries | 6 | 3 | 1 | 2 | Brad Flora frames the core dilemma for AI startups targeting legacy sectors and pushes Gustav on how founders can avoid getting bogged down in manual operations. Gustav provides a practical three-path framework and Airbnb metrics, maintaining a collaborative and constructive peer dynamic throughout. | |
| Enterprise vs. Mid-Market Targeting & Pace of Learning | 7 | 4 | 2 | 3 | Brad opens with a strong comparison between selling enterprise AI and high-risk moonshots, advocating for mid-market entry to increase learning velocity. Nicolas nuances the perspective by pointing out that certain enterprise pain points do not exist in mid-market or startup tiers. | |
| AI Employees vs. Founder Sales Execution | 6 | 2 | 1 | 1 | The panel is fully aligned that early founders cannot outsource core sales discovery to AI SDRs. Brad reinforces the YC canon that automation only scales a validated, working sales playbook rather than discovering one from scratch. | |
| Capital Allocation: Spending Now vs. Waiting for Model Leaps | 7 | 3 | 1 | 2 | The partners dissect whether to burn capital early or wait for base model upgrades, transitioning into deep discussion on when to pivot with modest traction. Brad contributes a direct case study from Greptile on user valuation versus vanity growth metrics. | |
| Y Combinator Application Call-to-Action | 5 | 2 | 2 | 2 | The segment includes a brief batch application callout followed by a debate on distinguishing 'good' from 'great' startup ideas. Brad introduces the provocative view that purely 'good' ideas are effectively bad because they trap founders in subscale outcomes. | |
| Navigating High Technical Difficulty in Startups | 6 | 3 | 2 | 2 | Nicolas counters conventional hesitation around high technical hurdles, asserting that extreme difficulty creates defensible moats. Brad details how he incrementally solved real-time ad bidding at Perfect Audience using modular API wrappers. | |
| Key Indicators for Hiring Your First Employees | 6 | 2 | 1 | 2 | The discussion focuses on warning founders against hiring too early or treating headcount as a prestige metric. Brad and Pete sharply narrow the scope of acceptable early hires to rare, outstanding opportunistic talent rather than prestige resume additions. | |
| Open Source Strategy for Enterprise SaaS & AI | 6 | 4 | 1 | 1 | Nicolas explains how open source functions as an enterprise trust and compliance accelerator rather than just a developer distribution channel. Brad notes the shift in enterprise willingness to accommodate on-prem and self-hosted AI setups. |
Statements from this episode (23)
Airbnb maintained a minimum technical staff ratio to prevent stalled productivity
“At Airbnb we had this metric, which was percent of technical people that work at the company, and the reason we had the metric is, at some point you have too many non-technical people all they do is request things from the technical people, and then you can't …”
Series A investors prioritize automation rate trajectory over overall revenue
“And if I was to say a CSA investor, and that would mean one of these companies, I actually care more about the trajectory of the automation rate than the overall revenue.”
Non-legal founders built Vessence MVP by embedding inside a law firm
“I worked with a company in the previous batch, the Spring batch, which was just now wrapping up, called Vessence, that are building software for lawyers. And neither of the founders has a legal background, and so the way that they got started before YC was the…”
Early B2B startups should target the smallest companies with their problem
“And probably You want to go after the smallest company that has the problem you try to solve.”
Finding an empowered buyer matters more than market segmentation in sales
“A lot of founders try to think of, like, segment before qualification, and I think sometimes Just the right person is more important as long as they are empowered.”
AI SDRs cannot rescue startup founders who are unable to sell
“AI SDRs tend to work well when they're plugged into a sales process that's already working well, and where I haven't seen them work well is when founders sort of turn to an AI SDR as the solution of last resort, where they're just totally unable to sell their …”
AI SDR startups targeting non-selling founders will face massive churn
“If they go after the people who are not able to sell their own product, there is a little chance that they can do better, and this Customers are going to churn. A lot of revenue fast, but mostly churn.”
Startup marketing VPs churn frequently because founders misunderstand the role
“VP of marketing is like notoriously high churn job because, and it's not because the VP of marketing folks aren't good. It's because the founders have the wrong expectations of what those people do, and they haven't been curious enough to learn about that job.”
Building startups solely to patch temporary AI model deficiencies is doomed
“If you are just building something that's solving for the pains that GPT-V is not yet solving, it's probably a bad idea.”
Investing ahead of future AI models justifies the temporary overspend
“If you do invest in it, if you do, you're going to learn a lot from the process, and once the models are going to be ready, you plug them, and your product is going to be much better day one. So indeed, you have maybe wasted, I don't think that's the right wor…”
Claude Sonnet instantly fixed many previously failing startup internal tools
“When that model came out, a lot of companies that were building, say internal tools were like suddenly working, and they weren't really working before.”
Firecrawl pivoted away from Mendable despite hundreds of thousands in ARR
“And I think that when they actually pivoted, they already had hundreds of thousands of dollars of AR. So, significant traction. Like, not like just you know, 200 bucks. Actually real customers. And big logos too.”
Pivoting startups should explore multiple ideas simultaneously to avoid demoralization
“It's much better when you're pivoting to have a range of different ideas exploring. So, sort of like, you can find conviction around something and be fine with throwing out a few of them.”
Losing conviction is the leading indicator that a founder should pivot
“The actual leading indicator that maybe you should pivot is you just stop believing that what you're working on is going to work out.”
Founders with truly great startup ideas rarely brag about having them
“Like, I have a great idea is not something that founders would have great ideas say generally.”
Extreme technical difficulty makes startup ideas better by deterring competitors
“If something is really hard,
On the technical side, I mean, I think that's an even better idea.
Like nobody else is going to try, right?
If it's hard, like the bar is so high, nobody tries, and nobody does it.
If you have the courage to actually do it, if you …”
Perfect Audience launched by wrapping a third-party real-time bidding API
“With my own company, Perfect Audience from years ago, we knew that we needed to build our own real-time bidding platform that was connected to all the ad exchanges and all these integrations, and it was very overwhelming.
We didn't even have any idea how to bu…”
Optimizely began as an internal bookmarklet used to deliver consulting contracts
“At Optimizely, the hardest part of building our first product was building this website editor.
That would work with any other website and would allow a non-technical person to go in and build an A-B test without writing code.
And that ended up taking us at le…”
If founders have free time to contemplate hiring, it is too early
“If you have a lot of time to think about this question, it's probably too early. If this is something that comes to mind every day for you, it's probably too early because It's the right time to hire when, like, things are so busy that you can't even find a sl…”
Hiring is an anti-failure measure rather than a startup success metric
“Hiring is not a success metric at all. It's sort of like a way to not go under or have a functioning company fail.”
Multiple Y Combinator startups aim for billion-dollar valuations with ten employees
“We have multiple Weiss companies who say we want to be a billion dollar, 10 person company, and we have six slots left.”
Early startup hiring slows founders down instead of speeding them up
“When I have founders that are working in the batch who ask if they should hire almost always the answer is no, right? Founders will make the mistake of thinking it will speed them up, but in reality it ends up doing the opposite.”
Open source models build enterprise trust and shorten regulated sales cycles
“I have that company in mind, Medplum, who is building an open source EHR, and I think for them, being open source was not about the go-to-market in the sense of, like, selling to developers. It was really about creating the trust at their customers, like, in t…”