20VC Newsletter - June 30th 2024
Here are the transcripts and top takeaways from 20VC episodes released this last week!
Monday’s episode with David Luan, Co-Founder @ Adept:
Download the full transcript:
My 6 key takeaways:
What OpenAI Realized That No One Else Did
The next phase of AI after transformers was not about research papers.
It is about identifying major unsolved scientific problems & trying to solve them.
It’s a totally different framework from the traditional academic approach.
Why I’m Not Worried About Diminishing Returns in Compute
There are 2 parts to model scaling with compute:
Make the model bigger & throw more data & GPUs at it.
Have the base model collect data & learn how to get smarter itself.
We’ve just started tapping into #2 & it will absorb a lot of compute.
Why Every Cloud Provider Must Have a Model Play
As models become smarter, they’ll become the base computing primitive.
The logic of software will be handled by LLMs in the future.
Whoever controls the model layer controls all of the underlying compute.
What is the Difference between RPA and Agents
RPAs are useful for high volume tasks that always look the same.
Agents are constantly planning & reevaluating every step to solve the goal.
People will be using computers for high-level goals in 5 - 10 years.
Why Pricing for Work is Not the Answer and is Co-Pilot for Incumbents?
Pricing by work is a corollary of AI taking all jobs & that’s not how it’ll play out.
Humans will fundamentally be the drivers of agentic systems that enhance their creativity.
This cannot be built without a co-pilot style approach.
Why Open Will Always Lag Closed
Open has fewer resources behind them.
There are fewer incentives for people to make things open as it gets more expensive.
I see it as a way for the rest of the fields to keep up with the incumbents.
Wednesday’s episode with Andrew Bialecki, CEO at Klaviyo:
Download the full transcript:
My 6 key takeaways:
This $6BN Company That Everyone Turned Down
We applied to several $20K no equity programs from VCs & everyone said no.
I told Ed if we couldn’t even get that how are we going to raise a $500K seed.
So we decided to do it ourselves.
Would Klaviyo Have Been Successful If it Had Raised VC Money Early?
Constraints breed a lot of creativity.
I used to jump between customer support & coding which really connected me to the customers.
My advice is don’t take more capital than you need.
It will distort the way that you build your business and interact with your customers.
The Road to $1M in ARR Ain’t Fricking Straight:
We started Klaviyo in 2012:
Year 1: $20K ARR
Year 2: $250K ARR
Year 3: $1M ARR
80% of the 36 months was just the two of us. You have to grind it out.
Why You Have to Go Slow to Go Fast
Having capital & moving fast is great if you have a plan.
BUT there is real magic in taking time to find product-market fit.
Building up customer love compounds and snowballs.
Why Partnerships Should Be a Tug of War
For a string to stay tight, you need two people pulling both ends of the rope.
The economic incentives need to align & we have that with Shopify.
The best partnerships are win-win for everyone.
Will AI Help SaaS Make More Money or Better Products?
There are 2 fundamental forces working in opposition:
All tech makes things easier, so there’s a deflationary element to it.
Building better tools also makes it increasingly valuable.
We haven’t put a price tag to the amount of increased output from AI.
Friday’s episode with Samir Vasavada, Co-Founder & CEO @ Vise:
Download the full transcript:
My 8 key takeaways:
I Was 20 When Our Company Was Valued at $1BN
Everyday a new billionaire or celebrity would ask to invest in my company.
It was a change in power dynamic.
It was like a child star phenomenon & it wasn’t healthy.
You Cannot Build a Remote Company
Culture can’t be built around Zoom calls all day.
It’s those in office water cooler moments that make a difference.
We were remote during the pandemic & our entire culture was off.
Why Hiring in the Boom Was So Broken
Executive recruiters would bring on 20 candidates with 25+ offers each.
They were all from big tech and ran 200+ person orgs.
We weren’t interviewing: We were convincing them to join our company.
People with Playbooks Cause Damage Pre-PMF
All the execs we hired came from companies with PMF.
They were maintaining the status quo & didn’t have to find PMF.
They used old playbooks & weren’t trying to understand our problems.
Why You Do Not Want to Hire the People Who Want to Join Cos Sequoia Did the Round
People thought Sequoia backed = rocket ship company.
You have to be extra mindful of why someone is joining your company.
Some people really believe in the mission and others just want an exit.
When There is Doubt There is No Doubt, Fire Fast
Too many founders & managers drag their feet on firing.
You have to keep your bar high as a startup.
You’re a basketball team & every player on the court matters.
The Problem with Hiring People with a Reputation
They are constantly trying to defend their reputation.
If something will remotely hurt it, they will flee.
They will jump the boat before you burn it.
99% of Advice is Bad Advice
And the rest of the advice is contradictory to each other.
Too many people assume advice from successful companies is good.
Our business wasn’t working until we stopped listening to everyone’s advice.
Let us know what your big takeaways from this week’s shows were in the comments below!
Thank you for reading, and don’t miss the great guests we have next week:
Matthew Clifford, Co-Founder @ Entrepreneur First
Avi Eyal, Co-Founder & Managing Partner @ Entrée Capital
For this week only, 20VC will take a break from our Friday show due to US public holidays. We'll be back with our full weekly schedule next week!