Why Effective Altruism Stopped Being Effective
Effective Altruism began with a deceptively simple premise: do the most good possible. This clarity of purpose drove early victories - malaria nets, deworming programs, direct cash transfers. Measurable interventions with quantifiable impact. The movement succeeded because it stayed close to observable reality.
Then something shifted.
EA's growth trajectory mirrors every successful movement: institutionalisation breeds complexity, complexity demands coordination, coordination requires consensus-building. What emerged was not more effectiveness, but more elaborate justifications for predetermined conclusions.
The telltale sign? When your methodology becomes more sophisticated than your results.
Modern EA discourse resembles academic philosophy more than emergency response. Thousands of hours spent modeling hypothetical AI scenarios while malaria kills a child every two minutes. The movement that once mocked "warm glow" giving now produces warm glow thinking - intellectually satisfying frameworks that optimise for coherence over consequence.
EA's Washington engagement reveals the fundamental tension. Movements that seek policy influence inevitably become shaped by policy constraints. The sharp edges get smoothed, the radical insights domesticated into acceptable recommendations.
Consider longtermism's trajectory: from "we should care about future generations" to "we should spend billions on speculative AI safety research." The philosophical insight remains sound. The policy translation became an exercise in motivated reasoning, where predetermined funding priorities discovered their philosophical justification post-hoc.
This isn't corruption.
It's convergent evolution. Organizations that successfully navigate political systems develop political immunities. They learn to speak in measured tones, hedge bold claims, build coalitions through compromise. These adaptive behaviors systematically select against the radical clarity that made EA effective.
EA's greatest strength became its limiting constraint. When everything must be quantified for comparison, only quantifiable interventions survive evaluation. The methodology that revealed global health's effectiveness now obscures interventions that resist measurement.
The paradox: optimization requires simplification, but complex systems resist simplification. When you optimise for what you can measure, you get more of what you can measure. Not necessarily more of what matters.
This manifests in EA's weird blind spots. Institutional reform receives minimal attention despite creating multiplicative effects across domains. Cultural and political entrepreneurship gets categorized as "too speculative" while longtermist cause areas receive serious consideration. The framework's precision masks its incompleteness.
EA optimized for scale before understanding sustainability. Rapid expansion diluted quality control. The community that once prided itself on changing minds through evidence became a social movement attracting adherents through tribal affiliation.
Network effects explain the deterioration: as community size increases, maintaining epistemic standards requires exponentially more effort. Ideas spread through social proof rather than rigorous evaluation. Popular conclusions get reinforced regardless of evidential support.
The result?
A movement nominally committed to truth-seeking that systematically produces motivated reasoning in service of community consensus.
Perhaps EA's deepest failure lies in its response to criticism. Instead of returning to first principles, the community generated increasingly sophisticated meta-frameworks to address concerns without changing behavior.
Criticism of longtermism's speculative nature? Develop better uncertainty quantification methods.
Concerns about funding concentration? Create new institutions with different names but similar leadership.
Questions about ideological capture? Commission studies on movement epistemics.
Each response adds layers of apparent sophistication while avoiding the core issue: the movement lost its connection to concrete impact measurement that originally justified its existence.
Effective Altruism can recover its effectiveness, but only through radical simplification.
First, return to outcome measurement over input optimization. Fund interventions based on demonstrated results, not theoretical promise. This means smaller experiments, faster iteration, lower tolerance for extended research phases.
Second, embrace institutional redundancy over efficiency. Multiple small organizations exploring different approaches outperform single large institutions optimizing existing frameworks. Competition preserves innovation; coordination preserves conformity.
Third, reconnect with ground-truth feedback loops. The most effective early EA organizations maintained direct contact with implementation reality. Distance from operational details enables theoretical drift.
The choice is stark: remain a philosophical movement that discusses effectiveness, or return to being an effective movement that occasionally discusses philosophy.
The world has urgent problems requiring immediate solutions. EA's original insight—that good intentions without effective methods waste precious resources, remains correct. But that insight now applies to EA itself.
Effectiveness demands effectiveness evaluation. By its own standards, current EA is ready for defunding.