Stay up to date on upcoming events, deadlines, news, and more by signing up for our newsletters.
Leslie Founders is a community of NYU student founders committed to launching and scaling ventures while at NYU. Members enjoy exclusive perks including today’s panel with Kenneth Auchenberg. Kenneth is a partner at AlleyCorp, a New York-based early stage venture capital firm, where he focuses on AI, developer tools, and infrastructure.
Before venture, Kenneth spent two decades building products and leading engineering teams. The conversation was not about AI hype. It was about what AI actually changes for founders.
AI Did Not Kill Entrepreneurship. It Compressed It.
Early in the discussion, a founder said something that anchored everything: the cost of code is collapsing, but the job of entrepreneurship is unchanged.
Startups are still about solving real problems for real customers. What changed is the timeline and the world.
What used to take six to nine months can now happen in weeks. In the pre-AI model, you raised a pre-seed, hired a team, disappeared, and shipped an MVP months later. Today, one or two builders with the right tooling can move at a pace that would have seemed unrealistic a few years ago.
The bottleneck is no longer writing software.
The bottleneck is judgment.
Feedback Loops Are the Advantage.
Kenneth emphasized iteration speed. The goal is no longer “ship when it is perfect.” It is “ship, learn, ship again.”
Talk to a customer and push an improvement within 24 hours. That sounds extreme until you realize how quickly teams can now prototype and deploy.
In a fast market, shipping two weeks earlier often matters more than shipping slightly better. No one remembers your first version. What compounds is learning velocity.
The Engineer Is Becoming an Orchestrator.
Engineering is shifting from pure implementation to system design and agent orchestration. Instead of writing every line, you are prompting, evaluating outputs, and exercising taste.
This creates tension for junior talent. If you do not know what good looks like, AI can make you wrong faster. Senior builders win because they recognize quality.
Enterprise Reality Still Wins.
The loudest AI conversations happen online. The largest budgets do not. Most enterprise buyers care about stability, compliance, and painful recurring problems. They are not chasing every new tool.
If you are building infrastructure, accept that you are building an enterprise business. Developers experiment. Enterprises pay.
Defensibility Is Moving Up the Stack.
When software becomes cheap, feature moat weakens. A motivated competitor can reach 80 percent parity quickly.
So what becomes defensible?
Distribution.
Brand.
Point of view.
In a world of infinite software, people choose products with taste and conviction. Opinionated workflows and identity can matter more than raw functionality.
From Services to Software, Again.
AI is turning parts of services back into software, unlocking vertical markets where work is repetitive and structured.
But the real asset is not just automation. It is the reasoning trace. The decisions and paths taken to produce outcomes. That data compounds.
Fundraising Follows Cost Structure.
If you only need one or two engineers and modest inference spend, early capital allocation shifts.
Less “raise to hire engineers.”
More “raise to win distribution.”
Capital still matters, especially for training models or unique datasets. But lean teams can reach product-market fit faster than ever.
The Bottom Line
AI is not replacing entrepreneurship. It is increasing the premium on what cannot be automated:
Customer understanding, judgment, taste, trust and distribution.
Code is abundant. Conviction and speed are scarce. If you are building now: the new path to seniority will be mastering system architecture and developing the judgment to evaluate agent outputs, not just writing code.
