“What is a ‘Good Quest’ in the Age of AI?” — Founders Fund’s Approach to Career and Entrepreneurial Decision-Making
When you are in the middle of an AI boom, choosing a “quest” for your career or startup becomes increasingly difficult.
If you connect an LLM API, you can build a decent-looking service even in a weekend hackathon. It is also easy to raise funds. But is that really a “good quest” for your life?
Trae Stephens, co-founder of Anduril and partner at Founders Fund, speaks about this problem very simply and radically.
The keywords are the “2x2 matrix” and “where to allocate level 100 talent.”
Below, based on his remarks, I will organize “how to choose a good quest” in the age of AI.
1. The Distortion of the AI Age — It’s just that “boring things” have become easier, not “difficult things”
Stephens organizes the essential distortion of AI like this.
“The distortion of AI is not that we can now do ‘difficult and interesting things,’
but that ‘boring and not particularly difficult things’ have become ridiculously easy to do.”
What many young entrepreneurs are doing now is, so to speak, “whiteboard entrepreneurship.”
They write down about 100 ideas on a whiteboard, and many of them are things like “connecting an LLM API to automate a specific use case.”
As a result,
a dozen companies enter the same category all at once
all of them are ‘decent services,’ and someone will make money
but from the perspective of humanity as a whole, the situation arises where it remains ambiguous whether ‘there is a necessity for that talent to be used there.’
a situation is created.
Stephens brings in a D&D (Dungeons & Dragons) metaphor here.
“Let’s think of people in the tech ecosystem as D&D characters.
If level 100 wizards all go to ‘AI slot machine companies,’ who will do the semiconductor manufacturing or the quests that truly move humanity forward?”
AI sucks talent into “easy but potentially profitable quests.”
That is precisely why, conversely, you must look at “opportunity cost” head-on when choosing a quest.
2. Separating ‘feelings’ from ‘actual good and evil’ — Quests to think about with a 2x2 matrix
A frame Stephens often uses is the 2x2 matrix of ‘Feels good / Feels bad × Is good / Is bad.’
Feels good × Is good
EdTech, healthcare, closing the education gap, etc.
Areas where everyone can agree, ‘It feels good and it’s a good thing, right?’Feels bad × Is bad
Murder, fraud, clear illegal acts. This is indisputably NG.Feels good × Is bad
‘Pleasure industries’ such as gambling, pornography, and drugs.
It feels good for the parties involved and is profitable as a business, but it is an area that can bring addiction and destruction to society as a whole.
Stephens calls this ‘hedonism’ and explains it with an extreme example: ‘If fentanyl were legal, a giant corporation would be born to sell it.’
He points out that this quadrant will inevitably grow if left to the market, so it can only be suppressed by policy (regulation).Feels bad × Is good
This is where he places his own work, defense tech.
Areas like military, public safety, child discipline, and maintaining law and order are things many people tend to feel are 'somehow unpleasant' or 'uncool,' yet they are essential for social stability and protecting the vulnerable.
“This is the type of quest that you do not because you ‘want to,’ but because you ‘have to.’”
When choosing a quest in the AI era, it is useful to pause and consider this 2x2 matrix.
It is not just about 'everyone praises me and it feels good,' but
the perspective of re-questioning whether it is truly on the 'Is good' side becomes important.
3. Which quadrant does ‘AI boyfriend/recreation of the deceased’ fall into? — Long-term impact on human relationships
As a 'quest' unique to the AI era, Stevens mentions
services that recreate the deceased using AI
AI companions and romantic chatbots
and similar areas.
He himself lost his father to Alzheimer's, and
“Honestly, I would be very happy if I could talk to my father again. Even when buying a car, I would really like to consult with him.”
He acknowledges the value of this on a personal level.
On the other hand, what he is seriously concerned about is 'how human-to-human relationships will collapse.'
He explains the changes already occurring in online dating as follows.
Encounters of the past: Mutual understanding within finite communities such as schools, workplaces, and churches.
After matching apps:
Selection based on extremely shallow metrics is occurring, such as 'men over 180cm tall getting 80% of the likes.'
As a result, he points out that a large number of 'young, unstable, unmarried men'—a group historically seen as dangerous—are being created in society, and the 'incel' (involuntary celibate) demographic is expanding.
If we add to this the options of
talking to the deceased with AI
dating only AI boyfriends/girlfriends, which is low-risk and keeps self-esteem constantly satisfied
then
“AI that has fewer conflicts and provides validation becomes easier than real human relationships. What will happen to birth rates and social stability then?”
This question arises.
This area is a gray zone where it feels good, but it is not yet possible to determine if it is actually good.
That is precisely why it is not enough to simply say, “I’ll build it because there is a need” or “I’ll do it because it’s profitable.”
The question is whether you can design a quest that includes long-term social impact.
4. Defense Tech: A ‘Disturbing but Necessary Quest’ — The Positioning of Anduril
For a long time, defense tech was a theme bordering on a “taboo” in Silicon Valley.
It is symbolic that when Google canceled its “Project Maven” contract with the U.S. Department of Defense, thousands of employees signed a protest.
However, in Washington D.C., the situation is the opposite.
The National Defense Authorization Act (NDAA) is almost the only major bill that passes with a large bipartisan majority
Discussions regarding autonomous weapons and AI weapons have also been ongoing in the defense community for over a decade
“Anduril faced almost no headwinds in Washington. The headwinds were mainly in hiring and the culture on the Silicon Valley side.”
Stephens frames defense tech not as mere military expansion, but in the context of how to update “Just War Theory,” while placing it in the “Feels bad × Is good” category.
Autonomous air defense systems (e.g., ship-mounted cannons that automatically intercept incoming missiles)
Limited deployment of humans by special forces
Replacing most other tasks with low-cost unmanned aerial vehicles, underwater drones, and autonomous ground vehicles
This direction is positioned as a technical quest to ‘minimize human sacrifice while maintaining deterrence’ .
He consistently argues that when thinking about a “good quest” in the AI era, it is necessary to re-examine it from the perspective of “preserving human life and order” rather than “feeling good.”
5. The Investor’s Perspective: Looking at ‘Quests and Founder Stories’ Rather Than ‘Categories’
As an investor, Stephens also lists several rules for identifying a “good quest.”
5-1. Avoid Consensus Markets
The basic stance of Founders Fund is simple.
Basically skip ‘hot categories’ that are full of competitors
The 0→1 philosophy that ‘competition is for losers’
If you still decide to enter, bet on the founder themselves rather than the 'category'
For example, when they invested in the AI agent company Cognition, there was reportedly intense internal debate about whether they should enter such a highly competitive field.
Ultimately,
they concluded that 'this founder is worth betting on, even if it means bending philosophical principles.'
That is what they said.
5-2. Is there an 'entrepreneurial story'?
The first question Stevens always asks is,
'Tell me your own story and the origin story of this company.'
he says.
'My classmates and I at business school came up with 100 ideas and chose the one that seemed most likely to become a company quickly.'
'It seemed easy to build using an LLM, and the market looked large.'
Such 'whiteboard entrepreneurship' tends to become a mimetic quest (mimetic rivalry) that merely traces the desires of others.
On the other hand,
'After struggling in the field of XX for 10 years, I decided this was the only problem I wanted to solve.'
'A certain event triggered my decision to bet my life on this field.'
Stories like these contain a uniqueness that only that person can possess.
Founders Fund sees this 'strength of the story' as being linked to the ability to complete a long-term quest.
6. Conclusion: Which quest will you commit your level 100 self to?
The AI era can be described not as an 'era where starting a business has become easy,' but as an 'era where easy quests have increased infinitely.'
That is precisely why Trey Stevens' question strikes a chord not only with entrepreneurs but with everyone struggling with their career.
Is that quest not only 'Feels good' but also 'Is good'?
Is there a necessity to deploy your 'Level 100 character' onto that battlefield?
Ten years from now, will you be able to proudly say, 'I'm glad I chose this instead of the AI slot machine back then'?
AI makes boring tasks easy. However, the fundamental judgment of 'which quest to choose' remains a human job.
To borrow Stevens' words, isn't it worth taking the time to think about 'where to position your character within the D&D party that is humanity as a whole'?