S2E7: "Be the Olympist of Your Field" | Guillaume Cabane

About this episode

Listen now | The founder of HyperGrowth Partners unpacks why hybrid GTM teams beat full automation, why true AI adoption stays under 1%, and why the next decade belongs to the Olympist of every field.

About our guest — Guillaume Cabane

Guillaume “G” Cabane is the founder and general partner of HyperGrowth Partners, where he and his team work hands-on with companies like Ramp, Neon, Zapier, AirOps, n8n, and Attention. He’s the recognized godfather of GTM engineering — pioneering the Clearbit reveal loop, personalized outbound, and the experimentation-as-system playbook before any of it was standard practice. Before HyperGrowth, G was VP Growth at Drift, Segment, and Gorgias, helping each one scale through multiple step-changes in revenue, and most recently served as interim CMO at Ramp through their run from $10M to over $1B in ARR.

Core Takeaways

* Hybrid GTM Beats Full Automation: G has yet to see a single scale-up successfully operate without a GTM team. He tested full automation against a hybrid setup (humans + AI) at Ramp himself, and failed to beat the hybrid baseline. Other people brought him their prompts and also failed. The AI-replaces-your-SDR-team posts are hype; SDR and BDR teams are still in active hiring across his entire portfolio.

* True AI Adoption Stays Under 1%: Most marketing teams have not done any meaningful AI work beyond a GPT prompt or two. When G builds something for them and shows the output, they’re happy with the output but don’t want to learn how to build it themselves. The technologists who actually compound value with AI are a tiny fraction — G estimates less than 1% of the population — and the gap between them and everyone else is going to widen every quarter.

* SaaS-for-SaaS Is the Risky Bet — Tech-for-Non-Tech Stays Safe: Build versus buy is shifting back to build. Niche subscriptions are getting harder to justify when you can vibe-code the same thing in an hour. SaaS sold to other software companies is the riskiest position — either LLMs start buying for them or enterprises start building in-house. Selling technology to non-tech companies (Ramp is the canonical example) stays safe; those buyers won’t suddenly start coding.

* Be the Olympist of Your Field: There won’t be hundreds of millions of Python developers in the future. There will be a few thousand master-level specialists who can beat the best LLM in their narrow domain. The path for younger talent is to pick a niche, dedicate ten years from age 15 to 25, and become the world-class expert. Most Olympians are out of college by 25. Why would tech be different.

Top Quotes

“I tried beating that hybrid setup with a full auto and I failed. And other people have come to me and gee, there must be a way to, we can improve the prompts. And they have failed.”

“True AI adoption, in the sense that you guys in this audience thinks about it, is going to be limited to a very small percentage of the population. I wouldn’t be surprised if it’s less than 1%.”

“I think SaaS for SaaS is extremely risky because either like LLMs are going to buy from that land or because large enterprise and mid markets are going to stop buying and like just start building.”

“You’ve got to be the Olympist of your field. That is how you survive. That is how you win.”

Referenced Tools and Resources

* AI & Building: Claude Code (referenced throughout as “Open Claude” and “Cloud Code”), GPT

* Workflow Automation: Zapier, n8n

* Content Production: AirOps

* Data Infrastructure: Salesforce (via MCP), vector databases

* Case Studies: Ramp (tech-for-non-tech archetype), Manifest (legal AI), Netic (applied AI in real-world spaces)

Timestamps

* (00:00) Cold open and intros

* (04:05) What G’s three-and-a-half years at Ramp changed about GTM engineering

* (06:45) What has and hasn’t changed about the growth team in 2026

* (08:30) The AI-replaced-our-GTM-team posts are hype

* (10:30) Hybrid wins, full automation loses — and G tested it himself

* (11:35) Is the GTM Engineer role being diluted, or is it the same person with a new title

* (14:30) Why true AI adoption stays under 1% of the population

* (17:45) The two societies splitting around AI and how fast the gap is widening

* (20:40) When everyone’s bot acts in their name, response rates collapse across every channel

* (24:30) The effective craziness framework — tech love plus rigor plus creativity plus psychology

* (28:10) What happens when LLMs start buying from LLMs

* (31:50) Have you given Claude Code your credit card yet

* (35:30) Are skills the new IP — and where defensibility goes

* (40:00) Build versus buy is shifting back to build

* (43:35) Why specialist agents still need specialists

* (46:30) 2026 is the year you either build or step off the progress train

* (50:30) SaaS for SaaS is the risky bet; tech for non-tech stays safe

* (52:55) Trust the model output instead of needing to know everything

* (54:30) Be the Olympist of your field

* (56:50) What G would tell college graduates today

Where to Find Guillaume

* LinkedIn

* HyperGrowth Partners

Where to Connect with Jared & Matteo

* Jared Waxman, GTM Engineer School Co-founder: LinkedIn

* Matteo Tittarelli, GTM Engineer School Co-founder: LinkedIn, X, Website, Newsletter

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit gtmengineerschool.substack.com