Pinned Tweet
The AI age demands a new kind of technologist: The Philosopher-Builder.
Every builder's first duty is philosophical: to decide what they should build for.
AI is beginning to decide what ideas reach your mind—your next action, your next job, your next relationship. It will tempt you to outsource your thinking in ways you’ve never been tempted before.
ecomes autocomplete for life.
200 years ago, when mechanization was starting to transform the world, the German philosopher Wilhelm von Humboldt warned of this very dynamic: To flourish as humans, we have to be self-directed. We are the architects of our own becoming.
But if we offload our thinking—if we let AI automate what we once pursued with intention—is it still our life to live?
To build AI that advances human flourishing, we need technologists who combine technical ability with moral vision.
Silicon Valley once embodied this: Jobs and Wozniak asked what kind of creative life personal computers should enable; ARPANET's pioneers envisioned what kind of connected society networks should foster.
But today, we now see three dominant archetypes:
1. The puzzle-absorbed: Solving technical challenges without reflecting on broader ends
2. The reductionists: Collapsing morality into math
3. The dismissers: Waving off philosophy as irrelevant to progress
All three abdicate the fundamental question: What should we build for?
This is where philosophy is essential. Because the job of a builder isn’t just to ship. It’s to shape a shared world.
We need a new kind of technologist—one who builds with human flourishing in mind. Who resists centralized control. Who strengthens autonomy instead of mining attention.
We call them Philosopher-Builders.
@cosmos_inst exists to train and support them—people who translate deep thinking into systems, institutions, and infrastructure that serve human ends.
If you believe the future of AI should expand human freedom and lift human potential, read the full essay.
5
19
127
39,955
Cosmos Institute retweeted
It’s easy to pit understanding against outputs in AI-for-science debates. But seeing understanding as an optional mental state revives old ideals of value-free science & mechanical objectivity that philosophers reject as insufficient.
Thanks @cosmos_inst for publishing this!
Thanks @cosmos_inst for publishing this!
“A paper whose authors can’t explain its claims leaves no one to consult if those claims fail.”
New guest essay from @JessicaHullman on why science needs people who can answer for results and challenge them, even once AI can verify them.
blog.cosmos-institute.org/p/verification…
1
5
18
2,160
“A paper whose authors can’t explain its claims leaves no one to consult if those claims fail.”
New guest essay from @JessicaHullman on why science needs people who can answer for results and challenge them, even once AI can verify them.
blog.cosmos-institute.org/p/verification…
New guest essay from @JessicaHullman on why science needs people who can answer for results and challenge them, even once AI can verify them.
blog.cosmos-institute.org/p/verification…
1
13
59
4,505
The idea of tail risk is nothing new - private markets have found ways of adapting and managing them in industries like nuclear and shipping.
@Brendan_McCord went on @MTSlive to explain how we could apply these lessons to AI.
@Brendan_McCord went on @MTSlive to explain how we could apply these lessons to AI.
1
3
9
1,164
Cosmos Institute retweeted
“Avoiding excessive deference to AI is important, but it is ultimately secondary to whether we can renew the virtues, traditions, and institutions required to form good choosers.”
@chalmermagne on why the AI deference debates we have are too narrow.
blog.cosmos-institute.org/p/beyond-freed…
@chalmermagne on why the AI deference debates we have are too narrow.
blog.cosmos-institute.org/p/beyond-freed…
1
7
36
2,956
Cosmos Institute retweeted
I’m hiring a Chief of Staff to be my right hand across everything I do at Cosmos.
That starts with Cosmos Institute (@cosmos_inst), the nonprofit I founded after selling two AI companies, having two kids, and thinking hard about the good life and the good society. Through the Institute, we do research internally and with fellows from @GoogleDeepMind, @AnthropicAI, @OpenAI, @Microsoft, and academia, have given out 200+ grants to build fast prototypes, and we’re launching a big new in-person bet for higher education in 2027. This side of my work also includes @UniofOxford, where I co-teach a graduate seminar on philosophy and AI and support the Human-Centered AI Lab (HAI Lab), as well as a book I’m writing on human autonomy in the age of AI.
On the venture side, you’ll work with me across Cosmos Ventures (@cosmos_vc), our new $77.6M fund backing founders building new institutions for the AI age. Our early portfolio includes companies like AIUC (@aiunderwriting), @PrimeIntellect, and Workshop Labs (acq. by @thinkymachines), with several more to be announced soon. We’re also going to be incubating new companies in 2027.
You’ll come in, do this well for two years, and if you want to build a thesis-aligned company or a nonprofit next, we will back you, or place you at a top company in our community. The role is based in Austin, with relocation provided.
To apply, read our essays and tell us where we’re wrong: docs.google.com/document/d/18_…
That starts with Cosmos Institute (@cosmos_inst), the nonprofit I founded after selling two AI companies, having two kids, and thinking hard about the good life and the good society. Through the Institute, we do research internally and with fellows from @GoogleDeepMind, @AnthropicAI, @OpenAI, @Microsoft, and academia, have given out 200+ grants to build fast prototypes, and we’re launching a big new in-person bet for higher education in 2027. This side of my work also includes @UniofOxford, where I co-teach a graduate seminar on philosophy and AI and support the Human-Centered AI Lab (HAI Lab), as well as a book I’m writing on human autonomy in the age of AI.
On the venture side, you’ll work with me across Cosmos Ventures (@cosmos_vc), our new $77.6M fund backing founders building new institutions for the AI age. Our early portfolio includes companies like AIUC (@aiunderwriting), @PrimeIntellect, and Workshop Labs (acq. by @thinkymachines), with several more to be announced soon. We’re also going to be incubating new companies in 2027.
You’ll come in, do this well for two years, and if you want to build a thesis-aligned company or a nonprofit next, we will back you, or place you at a top company in our community. The role is based in Austin, with relocation provided.
To apply, read our essays and tell us where we’re wrong: docs.google.com/document/d/18_…
19
56
474
63,447
RT @sethlazar: Announcing the Call for Papers for Oxford Studies in Philosophy of AI and Computing!
Rapid AI progress is raising *so many…
Rapid AI progress is raising *so many…
0
66
0
44
AI raises fundamental questions about intelligence, reasoning, and selfhood. Philosophy has a vital role to play.
We're supporting a new annual volume, Oxford Studies in Philosophy of AI and Computing, along with a $10,000 prize for its best paper.
blog.cosmos-institute.org/p/win-the-cosm…
We're supporting a new annual volume, Oxford Studies in Philosophy of AI and Computing, along with a $10,000 prize for its best paper.
blog.cosmos-institute.org/p/win-the-cosm…
1
29
133
7,626
Cosmos Institute retweeted
I often find debates around AI deference frustratingly narrow.
This is probably because they're grounded in philosophical accounts of autonomy that tie themselves in knots trying to establish when our choices are truly our own.
Here's my attempt at an alternative framing.
This is probably because they're grounded in philosophical accounts of autonomy that tie themselves in knots trying to establish when our choices are truly our own.
Here's my attempt at an alternative framing.
“Avoiding excessive deference to AI is important, but it is ultimately secondary to whether we can renew the virtues, traditions, and institutions required to form good choosers.”
@chalmermagne on why the AI deference debates we have are too narrow.
blog.cosmos-institute.org/p/beyond-freed…
0
3
15
775
“Avoiding excessive deference to AI is important, but it is ultimately secondary to whether we can renew the virtues, traditions, and institutions required to form good choosers.”
@chalmermagne on why the AI deference debates we have are too narrow.
blog.cosmos-institute.org/p/beyond-freed…
@chalmermagne on why the AI deference debates we have are too narrow.
blog.cosmos-institute.org/p/beyond-freed…
1
7
36
2,956
Cosmos Institute retweeted
Wrote a quick essay for this; feedback/comments/questions welcome: angadh.com/ai-and-educati…
AI agents could make organizations far more productive, but leave people more isolated and less autonomous.
We've teamed up with @inherent_labs on an essay prize exploring how organizations can harness AI while strengthening human autonomy and fulfillment.
We're thrilled that divyasiddarth, and Chris Summerfield are joining us as judges.
blog.cosmos-institute.org/p/essay-contes…
3
4
44
13,723
What is a philosopher-builder?
A philosopher-builder is someone who starts from a thesis about human flourishing.
Rather than simply looking for a gap in the market, they start out with a vision of the human capacities that are worth developing, what constrains them, and how hey reason from first principles about human potential and translate those insights into products or institutions.
To understand what a philosopher-builder is, it’s important to grasp what they are not:
1. They are not philosopher-kings. They do not have a blueprint for a society or a theory about how everyone out to live. Some of our most importance capacities are independent judgment, creativity, and the pursuit of knowledge. They seek to expand these capabilities, not prescribe outcomes
2. Open to disconfirming evidence. A strong philosophical starting point is open to rational criticism and has to be tested against reality. Their hypothesis is open to challenge and evidence should not only inform how they build their product, but potentially cause them to revise their underlying thesis. Philosophical seriousness demands intellectual humility.
3. Not a nihilistic iterator. “Make something people want” is a powerful entrepreneurial discipline, but by itself, it cannot establish whether satisfying people’s immediate wants advances human flourishing. You cannot outsource your belief in which ends are worth pursuing to market feedback.
The philosopher-builder is someone who combines philosophical conviction with empirical discipline. They have a vision, ambition, and an openness to revision.
We believe that AI isn’t a normal technology like the internet or the smartphone that society will simply absorb. Instead, it will create the biggest opening for institution-building since the American founding and philosopher-builders will be more important than ever.
If this vision of how to build resonates with you, we want to hear from you.
A philosopher-builder is someone who starts from a thesis about human flourishing.
Rather than simply looking for a gap in the market, they start out with a vision of the human capacities that are worth developing, what constrains them, and how hey reason from first principles about human potential and translate those insights into products or institutions.
To understand what a philosopher-builder is, it’s important to grasp what they are not:
1. They are not philosopher-kings. They do not have a blueprint for a society or a theory about how everyone out to live. Some of our most importance capacities are independent judgment, creativity, and the pursuit of knowledge. They seek to expand these capabilities, not prescribe outcomes
2. Open to disconfirming evidence. A strong philosophical starting point is open to rational criticism and has to be tested against reality. Their hypothesis is open to challenge and evidence should not only inform how they build their product, but potentially cause them to revise their underlying thesis. Philosophical seriousness demands intellectual humility.
3. Not a nihilistic iterator. “Make something people want” is a powerful entrepreneurial discipline, but by itself, it cannot establish whether satisfying people’s immediate wants advances human flourishing. You cannot outsource your belief in which ends are worth pursuing to market feedback.
The philosopher-builder is someone who combines philosophical conviction with empirical discipline. They have a vision, ambition, and an openness to revision.
We believe that AI isn’t a normal technology like the internet or the smartphone that society will simply absorb. Instead, it will create the biggest opening for institution-building since the American founding and philosopher-builders will be more important than ever.
If this vision of how to build resonates with you, we want to hear from you.
16
26
264
7,429
Cosmos Institute retweeted
"We live in an age of maximum information but minimum reflection."
New guest essay from Devin Stauffer, Professor of Government at @UTAustin and Cosmos Founding Fellow on why reading the Great Books is more urgent than ever.
blog.cosmos-institute.org/p/reading-the-…
0
3
23
807
"Those are counsels of passivity at a time when we need founders. The people in our community are not like that."
@Brendan_McCord on @MTSlive, discussing why the AI age needs new Franklins like @luke_drago_, @vincentweisser, and @RuneKvist.
@Brendan_McCord on @MTSlive, discussing why the AI age needs new Franklins like @luke_drago_, @vincentweisser, and @RuneKvist.
1
0
28
887
Cosmos Institute retweeted
Submit your essay! I look forward to reading it.
AI agents could make organizations far more productive, but leave people more isolated and less autonomous.
We've teamed up with @inherent_labs on an essay prize exploring how organizations can harness AI while strengthening human autonomy and fulfillment.
We're thrilled that divyasiddarth, and Chris Summerfield are joining us as judges.
blog.cosmos-institute.org/p/essay-contes…
2
9
56
5,960
AI agents could make organizations far more productive, but leave people more isolated and less autonomous.
We've teamed up with @inherent_labs on an essay prize exploring how organizations can harness AI while strengthening human autonomy and fulfillment.
We're thrilled that divyasiddarth, and Chris Summerfield are joining us as judges.
blog.cosmos-institute.org/p/essay-contes…
We've teamed up with @inherent_labs on an essay prize exploring how organizations can harness AI while strengthening human autonomy and fulfillment.
We're thrilled that divyasiddarth, and Chris Summerfield are joining us as judges.
blog.cosmos-institute.org/p/essay-contes…
2
30
177
26,491
Cosmos Institute retweeted
Introducing SnapJudge: a free, open AI detector for scientific writing, including math!
Drop in a PDF screenshot, paste text or LaTeX, or just put in an arXiv link.
We ran it on arXiv: ~1 in 3 new math papers are flagged as mostly AI, up from ~0 in early 2025. CS is 45%.
🧵
Drop in a PDF screenshot, paste text or LaTeX, or just put in an arXiv link.
We ran it on arXiv: ~1 in 3 new math papers are flagged as mostly AI, up from ~0 in early 2025. CS is 45%.
🧵
9
17
77
12,102
Cosmos Institute retweeted
New piece by myself and @lawhsw on what philosophy of science can tell us about AI development. Models are currently great at verifiable tasks, because they start out with a bunch of relevant context and domain knowledge, and can Try Stuff Until It Works; then we can reinforce that Stuff till they're extremely good. But in most areas, it isn't easy to check if It Works. So will the success of models be isolated to verifiable domains?
We argue not. Becoming competent in any area requires two things: a map of the relevant territory, and an ability to navigate it. Pretraining teaches the former, and RL teaches the latter. Why expect RL to transfer? As Popper argued, the basic pattern of navigation — proposing ideas, finding faults, trying alternatives — is common to all endeavours involving the growth of knowledge.
So training on maths and code doesn't just teach models maths and code. It teaches reusable ways of searching for procedures that transform an initial state into a desired one. Getting from a blank Lean document to a completed proof obviously requires different procedures from turning a rough draft into a great essay. But the abstractions learned in discovering the former make the latter easier to learn, provided pretraining has given the model enough knowledge of the new domain.
"A verifier determines what earns a reward; it underdetermines what a model learns in trying to earn it". Or, verifiability tells us where capabilities are easiest to train, not the limits of what that training can teach.
Read it! blog.cosmos-institute.org/p/verification…
We argue not. Becoming competent in any area requires two things: a map of the relevant territory, and an ability to navigate it. Pretraining teaches the former, and RL teaches the latter. Why expect RL to transfer? As Popper argued, the basic pattern of navigation — proposing ideas, finding faults, trying alternatives — is common to all endeavours involving the growth of knowledge.
So training on maths and code doesn't just teach models maths and code. It teaches reusable ways of searching for procedures that transform an initial state into a desired one. Getting from a blank Lean document to a completed proof obviously requires different procedures from turning a rough draft into a great essay. But the abstractions learned in discovering the former make the latter easier to learn, provided pretraining has given the model enough knowledge of the new domain.
"A verifier determines what earns a reward; it underdetermines what a model learns in trying to earn it". Or, verifiability tells us where capabilities are easiest to train, not the limits of what that training can teach.
Read it! blog.cosmos-institute.org/p/verification…
"Judgment can be informed without being mechanical: editors are not guessing blind."
@_AashishReddy and @lawhsw argue that skills obtained in verifiable domains like math and code carry over to unverifiable domains like writing.
blog.cosmos-institute.org/p/verification…
2
7
60
8,020
"Judgment can be informed without being mechanical: editors are not guessing blind."
@_AashishReddy and @lawhsw argue that skills obtained in verifiable domains like math and code carry over to unverifiable domains like writing.
blog.cosmos-institute.org/p/verification…
@_AashishReddy and @lawhsw argue that skills obtained in verifiable domains like math and code carry over to unverifiable domains like writing.
blog.cosmos-institute.org/p/verification…
0
5
40
4,388
"No one who advocates for pacing can really tell me what power the pacer would not have."
@Brendan_McCord went on @MTSlive to talk about how calls to 'pace the frontier' risk turning into a blank check for government to reach further and further into frontier labs.
@Brendan_McCord went on @MTSlive to talk about how calls to 'pace the frontier' risk turning into a blank check for government to reach further and further into frontier labs.
1
1
14
893