Most hard tech startups do not fail because they are too ambitious.
They fail because they choose the wrong kind of hard.
Shahin Farshchi, General Partner at Lux Capital, has spent 18 years backing companies like Zoox, Relativity Space, and Mosaic.
His core insight is deceptively simple. Not all hard problems are the same.
Shahin breaks “hard” into two very different buckets.
Scientific risk → Fundamental discoveries that do not exist yet. His view is that this belongs in academia.
Technical difficulty → Engineering known science into a real, defensible, monopolistic product.
The mistake founders make is not aiming high.
It is underwriting discovery when the job is execution.
What’s Inside the Episode
In this week’s episode of The Library of Minds, Dara and Shahin unpack how elite deep tech investors think about risk before metrics exist.
The conversation covers:
The path from PhD to founder and choosing hard problems wisely
The “ask nothing” strategy for breaking into venture capital
How to balance technical difficulty versus scientific risk
Why deep tech startups often only get one real shot
The controversial Mosaic investment and what it revealed
Learning to trust the bets you have already made
Hard truths about scaling from 1 to N
Why money is the ultimate commodity
To understand why choosing the right kind of hard matters more than choosing the hardest problem, you have to hear how Shahin connects it all.
The Core Distinction
Scientific breakthroughs create knowledge.
Technical breakthroughs create companies.
Shahin believes the best hard tech founders know the difference. They build on what is already true, then execute relentlessly to turn it into something only they can own.
That clarity is what separates enduring deep tech companies from ambitious experiments that never escape the lab.
Step Inside Shahin’s Mind
Shahin has created a Digital Mind on Delphi where you can ask him questions directly and hear his answers in his own voice.
Watch the Episode and Talk to Shahin’s Delphi
Try asking:
How do I know if my startup is taking scientific risk or technical risk?
What makes a deep tech bet investable before revenue exists?
When should founders walk away from a problem that is too early?
Delphi
Open your mind.
