Org-builder focused on philosophy of technology and education reform.
Director of the Institute for Classical Dialogue [Serving 100+ underserved teenagers in New Mexico with literacy-based enrichment programs]
Publisher of the “Better Argument Magazine” [Supporting 30+ writers working on philosophy of technology and other under-developed fields of existential philosophy.]
Hartree Consulting [Helping mission-driven orgs—redteaming, stategic consulting.]
Please forward this to anyone working in a junior rule in AI safety/governance orgs.
Direct call for articles on the following:
[visit better-argument.com/writers for details]
Economics & Political Economy
What happens when everyone says it’s a bubble?
How is AI-driven job replacement economically sustainable if displaced workers are also the consumers?
Is “productivity” still a coherent measure of economic value in an AI-mediated economy?
Are we moving toward an economy of ownership (compute, data, models) rather than labor?
What replaces advertising if persuasion becomes too powerful?
Technology, Platforms & Incentives
What would a social platform look like that is both economically viable and aligned with human flourishing?
What would it mean to design a platform that optimizes for wisdom rather than engagement?
What institutional structures could support slow thinking in fast media environments?
Is virality fundamentally at odds with truth?
Education & Knowledge
Should education train memory, judgment, or tool use—and in what proportions?
What becomes of the university if it no longer has a monopoly on knowledge transmission?
What replaces peer review when production becomes cheap and constant?
Can there still be a canon in an age of infinite content?
Work, Identity & Intellectual Life
What does authorship mean when content is co-produced with machines?
Is there a future for generalists, or only coordinators of machines?
What does intellectual honesty look like when assistance is invisible and universal?
Does the concept of vocation survive technological displacement?
Governance & Power
What forms of governance are appropriate for technologies that evolve faster than law?
Can states regulate systems that are globally distributed and privately owned?
Should highly capable AI systems be treated as infrastructure, firms, or something else?
Is centralized control of AI development stabilizing or dangerous?
Ethics & Human Meaning
How should we understand dignity in a post-work society?
Is it wrong to replace human relationships with artificial ones?
What ethical frameworks are robust to rapid, discontinuous change?
When do historical analogies clarify, and when do they mislead?
Culture & Society
What forms of culture emerge when audiences can generate exactly what they want?
Thankyou for your post! Would you please detail more about how you landed your first role?
Thankyou! Can you suggest anyone that I could talk to at GWCC who works on evaluating evaluators? I would imagine someone like this would have a lot to say about the best methodologies.
I’m very interested to learn more about testing. This would be especially important when I am at the stage of reaching churches and Christian businesses to offer auditing based on charity evaluations. First I need to develop a framework for evals.
This is a helpful insight about effectiveness vs. cause prioritisation. Especially in Christian circles (anecdotally) there’s a high amount of resistance to drop an existing effort for a more important one.
Hi, the link to fellowships is currently broken. Can you please fix it?
This intersects with a common criticism against EAs: that utilitarianism doesn’t work and is very silly, because of such problems like eating tuna fish to prevent tunas eating other things. Although maybe this just points out that utilitarian value calculations are just extremely hard, not wrongheaded.
I would describe myself as a conservative EA.
I think I got to a conservative political compass because of cause prioritisation. I felt that Liberal political groups were focused on personal identity considerations at the expense of more important goals. An example would be diversity hiring over talent hiring (which might be an overblown, optical concern that doesn’t really exist as much as I think it does.)
My feeling is that dividing up identity groups based on what makes one similar (similar gender identity or sexual orientation) is the wrong way of understanding community. Rather community is predicated on the need for diversity (builder, baker, candlestick makers make community. Not straight men.) I am an EA because of my similarities with EAs, but I am a member of my local community first and foremost. That is the community I need to exist, local community is more important than identity based communities like EA. [I really welcome red-teaming on this!]
I think EA’s have the mental strength to handle diverse political views well. What makes it difficult for conservatives and liberals to have conversations is an unwillingness to view others complexly, and being unwilling to assume good intent. We are EA’s because we care about doing good well. We already assume a certain amount of good intention in interaction with other EA’s. We could signal this to conservative folks. Liberals do not have a monopoly on moral feelings.
That said, a real conservative/liberal divide is the size of one’s circle of moral concern. EA’s have very broad ones, conservatives have very local ones. But this also seems like a general problem with EA, that it can think too large, too broad, too un-local. I’m pretty unsure about that.
Some major weaknesses of this post worth admitting to:
A major critique of “formation-based” solutions is that they are notoriously “thick” and hard to scale. It’s easy to scale a Universal Basic Income (UBI) algorithm; it is very hard to scale a Socratic mentor or a master craftsman.
If meaning-making requires “communities of excellence,” we face a massive talent constraint. I haven’t yet solved the question of how we produce enough “formative mentors” to meet the needs of billions of displaced workers without the quality collapsing into the “performative bureaucracy” I warned about in the Goodhart’s Law section.
One might ask: How is this different from simply funding more humanities departments? The distinction I’m trying to draw is that current humanities education is often optimized for credentialing (becoming a “worker”), whereas I am arguing for habituation (becoming a “human”). However, I concede that the “institutional form” this takes—whether guilds, leagues, or seminars—remains under-defined.
I’m particularly interested in hearing from people working in Cause Prioritization: If we accept that AI will eventually solve the “production” side of the human equation, how much should we be spending now on the “meaning-formation” side? Is this a neglected cause area, or a “soft” problem that we shouldn’t touch until Alignment is solved?