Good morning 👋
Today we’re breaking down a conversation with Kevin Novak (novakkm) (GP @ Rackhouse).
Rackhouse focuses on founders at the intersection of AI/ML and the real world, and the goal according to their website is to help the most competent and least discovered AI founders win.
I love that - might start copying it.
Kevin is one of the smarter AI investors that I follow, and I wanted to get his take on where he sees things going from his seat.
Let’s get into it.
P.S. If you’re looking for more investor interviews like this one …
Kevin Novak
GP @ Rackhouse
Key lessons from time at Uber and thoughts on people getting big tech experience before striking out on their own
“Advice usually falls into three buckets: do the good things faster, ignore the irrelevant stuff you're overthinking, and don't miss the one killer risk you're not paying attention to.”
“At Uber, we were told: you don’t recognize how special this is—you need to be going even faster. That taught me how to spot moments that matter and to sprint through them.”
“The stuff that blew up in our face? It was usually because we reinvented something we shouldn’t have—like HR. Don’t reinvent mission-critical functions.”
Founder errors of the past vs. founder errors in today’s world
“Ten years ago, the big mistake was over-architecting pre-PMF. Today, it’s falling in love with architecture and refusing to evolve. If you're not re-architecting, you're falling behind.”
“You got left behind three OpenAI press releases ago.”
Systems of record and changing data moats
“What used to be considered secret sauce—like building your own vector database stack—is now available via API. The moat isn't your model anymore; it's your infrastructure and your data layer.”
Second-order effects of intelligence becoming cheaper
“Replacing $100K junior hires with $20/month AI tools makes sense today. But the real winners will be firms that combine analysts and AI.”
“We're hitting parity with mediocre customer support. The next wave will deliver experiences that are actually better than what existed before.”
“As expectations rise, roles will evolve. You might be managing a Zapier pipeline and covering demo days."
New business models created from AI-first businesses
“Trying to predict the next AI business model is a trap. We’re not spearfishing—we’re hunting. We focus on coverage and speed.”
“Cloud-native beats cloud-adapted. Taxi Magic took taxis and added tech. Uber reimagined the system from scratch.”
How do you think about investment time horizons especially today when things are changing as quickly as they are today?
“You can’t thesis your way into a market like this. You have to stay horizontal early, then specialize once patterns emerge.”
“Benchmarks used to say, 'We invest in computer networks.' a16z started with 'software that lives on the internet.' Be broad early and sharpen later.”
Why infrastructure will matter more than algorithms
“Compute, storage, and sensors are the real unlock. A routine data scientist today has more power than I did as a supercomputing specialist.”
“You need a substrate before you can build models. Cloud infrastructure gave us that.”
What is the worst part about venture that nobody talks about?
“Honestly? Other VCs. It’s transactional and isolating. A lot of people you’d never hang out with outside work.”
“You say no 99% of the time. That wears on you.”
Who is another emerging GP who has impressed you?
“Nicole Wishoff. She’s amazing at content, wins deals, and raised a fund while pregnant. I’m jealous.”
TWEET
Thanks for reading this far and giving us a little bit of your attention this week.
Feel free to unsubscribe whenever this stops becoming valuable to you.
- Clay
"This has had a 100x ROI."
See why some readers are bumping up their reading experience ....
Upgrade todaySee which plan is right for you:
- Insider-only posts
- Operator-level access to give you an edge
- Boardroom-level support from our team (whenever you need it)
