The Twenty Minute VC (20VC) – "Why 90% of Founders Build Startups Wrong | Why AI Growth Rates are Sustainable & Remote Work is BS and the AI Talent War"

Guest: Jesse Zhang (Co-founder & CEO, Decagon)
Host: Harry Stebbings
Date: September 19, 2025

Episode Overview

In this high-energy Memo episode, Harry Stebbings sits down with Jesse Zhang, co-founder and CEO of Decagon, a fast-growing conversational AI platform for customer experience. The conversation dives deep into Jesse’s founder journey, scaling with AI, navigating hypergrowth, the realities of enterprise AI sales, and the “war for AI talent.” Jesse offers tactically dense, candid insights on building startups the right way (and the wrong way), why most founders fail, and what really matters when competing in one of the hottest markets today.

Key Discussion Points & Insights

1. From Math Olympiad to Startup CEO

• Correlation between math competitions and startup success

  • Jesse highlights the reasoning skills and ambition fostered in Math Olympiads as a template for effective founders.
  • “It just turns out that these folks do very well in startups.” (05:04, Jesse)

• Idea: Fund exclusively Math Olympiad founders

  • Lighthearted agreement between Jesse and Harry on starting a fund focused on Math Olympiad alums, hinting at the deeper topic of founder selection bias. (05:43)

2. Lessons from First Startup & Avoiding Common Mistakes

• Early mistakes: Over-intellectualization

  • Jesse admits his biggest mistake was overthinking markets and trends instead of talking to customers.
  • “We just over-intellectualized it. We built things people didn’t really care about.” (06:45, Jesse)

• The power of execution over market selection

  • In early stages, testing and discovery trumps picking the “right” market up front.
  • “Execution should help you find the right markets because if you’re actually executing discovery well, you should be able to discover, okay, which markets actually will be real versus not.” (09:56, Jesse)

3. How to Build a Startup in 2025—Jesse’s Approach

• Customer-led discovery > Noise and Hype

  • Ignore tech press and “hot trends” — focus on direct customer conversations and willingness to pay.
  • “We just spent a ton of time talking to customers directly…what’s useful to you? How much would you pay for this?” (08:06, Jesse)

• Zero to $1M ARR in Six Months

  • Achieved with just two founders, by moving fast and iterating tightly with customer feedback.
  • “Biggest learning was we really just went deep on discovery.” (13:16, Jesse)

4. Brand Name Venture Investors: A Double-Edged Sword

• Talent & customer signal, but no platform magic

  • Brand VCs help attract talent and give credibility, but can't accelerate you to product-market fit.
  • “It’s basically impossible for any VC to help accelerate the process of getting to PMF.” (15:21, Jesse)

• On signaling risk:

  • Not a worry for Decagon due to high demand, and Jesse prefers different VCs at Seed and Series A. (16:13)

5. AI Disruption: Starting Fresh Beats Legacy Integration

• AI-native startups vs. incumbent baggage

  • Incumbents are handicapped by legacy systems and customer contracts, making true AI transformation easier for startups.
  • “Having the older way of doing things serves as baggage.” (17:15, Jesse)

• Decagon's core innovation: AOPs (Agent Operating Procedures)

  • Allows non-technical users to build and iterate with natural language, unlocking broader adoption in CX. (18:03)

6. AI Value Capture: From Software Spend to Human Labor Budgets

• Decagon’s market grows by tapping into labor cost savings.

  • Deals are justified by potential labor replacement, unlocking much bigger budgets.
  • “Human labor is generally like an order of magnitude larger than software spend—like 10x or more.” (22:23, Jesse)

• Minimal resistance from CFOs:

  • AI is already a top-down initiative in most companies; the key is proving real ROI. (23:11)

7. Going Beyond Customer Support: The Future of Conversational AI

• Expanding the scope of AI agents

  • Moving from reactive support to proactive brand interfaces, covering more of the customer lifecycle.
  • “It becomes almost like a new conversational UI of your brand.” (24:32, Jesse)

• Switching costs in enterprise AI

  • True lock-in comes from unique logic and intelligence, not just data—which Decagon tries to make transparent and portable. (39:47, Jesse)

8. Competitive Landscape & Growing Pains

• Jesse’s most respected competitors:

  • Salesforce (for scale) and Sierra (for execution speed), while noting differences in approach (horizontal vs. vertical, configuration-heavy vs. product-forward). (34:51, Jesse)

• On high valuations:

  • Jesse intentionally avoided raising at peak multiples to ensure motivation and future flexibility.
  • “It just creates, like, okay, now the goalpost is here…all the wins you do have feel a little bit diluted.” (37:05, Jesse)

9. The AI Talent War & Remote Work is “BS”

• In-person, high clock-speed teams win

  • Prefers in-office for productivity and culture, although he acknowledges remote had its place during COVID. (33:18, Jesse)
  • “We value clock speed really, really highly…how fast your brain works and how fast you can learn.” (41:36, Jesse)

• Hiring in AI is the biggest challenge, but Decagon’s culture resonates with ambitious, hard-working talent

  • “At our stage we cannot get enough engineers.” (32:08, Jesse)

10. Culture, Stress, and Winning

• On stress:

  • Modern founders are too focused on wellness; embracing stress can be a strength.
  • “If there's no stress in your life, your life's not as exciting.” (46:13, Jesse)

• Building a winning culture:

  • “Managing stress is not what's important. It's: is what I’m doing meaningful? And do I get joy out of these milestones?”
  • Inspired by Nick at Revolut: “People want to be on a winning team.” (47:41, Harry quoting Nick/Revolut)

Notable Quotes & Memorable Moments

• On ignoring hype:
  "The number one thing that people should do is just ignore all of that. It's obviously useful signal, but it's just way too dangerous to read too much into that." — Jesse (08:06)

• On raising from VCs:
  "It’s basically impossible for any VC to help accelerate the process of getting to PMF.” — Jesse (15:21)

• On the AI agent value proposition:
  “If you’re building enough of a product, you’re more benchmarked against what is the business problem you’re solving. And the business problem you’re solving is going to be way bigger: saving human labor, transforming their customer experience, increasing revenue.” — Jesse (20:00)

• The importance of “clock speed”:
  “We value clock speed really, really highly, and this is across all functions... how fast your brain works and how fast you can learn.” — Jesse (41:36)

• On remote work:
  “It’s just way more productive [in person]... Ashwin and I are doing a lot of communication in person. It just allows ideas to flow a lot faster.” — Jesse (33:20)

• On stress and startup life:
  “People are just over indexing on this stuff... if you invest so much into wellness, it actually makes the stress worse because now there’s this huge juxtaposition between relaxing and stress. Instead, just embrace it.” — Jesse (46:13)

Important Timestamps

• 05:04 — Math Olympiad background and startup DNA
• 06:45 — Early founder mistakes: over-intellectualizing
• 08:06 — Execution vs. market selection, and the right way to build
• 13:16 — Zero to $1M ARR: discovery, speed, and focus
• 15:21 — VC brands and limits of platform value
• 17:15 — Why incumbents can’t keep up in AI
• 18:03 — Decagon’s AOPs: making AI accessible to non-engineers
• 22:23 — Unlocking spend: human labor budgets vs. software
• 24:32 — The customer support AI agent of the future
• 32:08 — Team growth and AI engineering demand
• 33:20 — Why Jesse prefers in-person teams
• 37:05 — The pitfalls of raising at excessive valuations
• 41:36 — “Clock speed” as the #1 hiring criterion
• 46:13 — Embracing stress vs. the modern “wellness” movement

"Quickfire" Segment Highlights (50:58–55:42)

• OpenAI at $300B or Anthropic at $60B?
  “Anthropic. That multiple will shrink over time. Anthropic is doing interesting things, especially with coding agents.” — Jesse (51:07)

• Most overhyped part of AI:
  “That AI is just going to transform every single use case. It’s not happening.” (52:01)

• If he could poach anything from a competitor:
  “The distribution of Salesforce. You got 40,000 reps just instantly going out there.” (52:27)

• Biggest internal debate at Decagon:
  “How to balance customer success/growth with investing in long-term product development and hiring.” (53:55)

• Exclusive model access or unlimited engineering hires?
  “The latter, for sure. No question. Most of the gains are from building around the models, not just the models themselves.” (54:59)

Final Reflections

• On Leadership Weakness:
  “It’s hard for me to escape the low-level details. I need to be more balanced and strategic.” — Jesse (55:55)

• On the future of AI in Sales:
  “In three to five years people will figure out some things where AI can be really useful [in sales].” — Jesse (50:00)

Summary Takeaways

1. Building a successful AI startup is about relentless customer focus, not intellectualizing market trends.
2. Incumbents are slow due to legacy “baggage”—this is the time for AI-native startups.
3. The real prize in enterprise AI isn’t just software budget, but tapping into massive human labor spend.
4. Brand name VCs help with talent and customer validation, but can’t deliver product-market fit.
5. Startups should embrace hard work, stress, and “clock speed” as virtues—remote is (for Jesse) less effective.
6. The future is not winner-take-all; multiple players will thrive in enterprise AI.
7. Engineering talent is the single biggest constraint—more than any technical advantage right now.

For those who haven’t listened: This episode is a masterclass in pragmatic startup execution, founder psychology, and AI industry realities—delivered in a rapid-fire, candid, and entertaining tone. Jesse Zhang’s clarity on what it actually takes to win (and what holds most founders back) makes this essential content for anyone interested in building, funding, or scaling category-defining startups in the AI era.