# When Benevolent Dictators Meet Apostolic Authority Single-point authority structures create remarkably similar conditions for harm whether they emerge in charismatic churches or open source software projects. The New Apostolic Reformation's apostolic governance model and the tech world's "Benevolent Dictator for Life" pattern share core structural vulnerabilities: both concentrate power in individuals who claim special vision, both lack meaningful accountability mechanisms, and both deploy psychological dynamics—loyalty rhetoric, critic-dismissal, departure narrative control—that silence dissent. The critical insight from cross-domain analysis is that **good intentions do not prevent these harms**—the structure itself creates conditions for abuse regardless of the leader's character. This parallel illuminates why Matt Mullenweg's 2024 WordPress crisis echoed patterns documented in NAR church scandals, why Guido van Rossum's Python burnout mirrors pastoral burnout, and why survivors of both contexts describe strikingly similar experiences of gaslighting and silencing. Both domains are now grappling with the same fundamental question: how do you preserve vision and efficiency while distributing power and ensuring accountability? --- ## The architecture of single-point authority Both NAR and BDFL models rest on a theological or philosophical claim: that certain individuals possess unique gifts qualifying them for supreme authority. In NAR churches, this manifests through the "five-fold ministry" structure derived from Ephesians 4:11, with apostles at the apex. C. Peter Wagner, the movement's founder, articulated the shift explicitly: "In traditional denominations, the locus of authority is ordinarily found in groups, not in individuals... **In the New Apostolic Reformation, however, trust has shifted from groups to individuals.**" The open source parallel is striking. As David Heinemeier Hansson (Rails creator) wrote defending the BDFL model: "Half the advantage of the BDFL model is exactly in allowing for unpopular decisions to be made without the lethargic mores of committees and bureaucracy!" Both frameworks explicitly reject distributed authority in favor of individual discernment. The authority structure operates through similar mechanisms. NAR apostles receive "prophetic vision" through a chain where intercessors pray, prophets receive revelation, and apostles implement God's will by "casting vision." Software BDFLs possess what one Facebook employee called "keeper of messianic vision"—an overwhelming picture of a new world that others must trust. In both cases, **vision functions as a trump card that inherently excludes dissent**, since questioning the vision means questioning either divine revelation or the founder's unique insight. The WordPress crisis illustrates how this plays out in practice. Matt Mullenweg simultaneously controls Automattic (the for-profit company), the WordPress Foundation (nonprofit), and WordPress.org (community infrastructure)—a concentration of power mirroring how NAR apostles often control church, missions organization, and prophetic network simultaneously. When Mullenweg attacked WP Engine as a "cancer to WordPress" and blocked their access to WordPress.org, he wielded infrastructure power against a competitor in ways that harmed over a million uninvolved users. The structure enabled harm regardless of the legitimacy of his underlying concerns. --- ## Missing accountability mechanisms expose identical vulnerabilities The structural gap most consistently identified across both contexts is the absence of meaningful accountability. As the academic journal _Spiritus_ noted about NAR: "To date, there are **no valuable, concrete, detailed case studies in the academic literature about the new movement's actual leadership practice**"—a documentation gap that itself reflects the movement's resistance to external scrutiny. NAR "covering" theology explicitly subordinates believers to apostolic authority. According to this framework, "rebellion against God's delegated authority is rebellion against God" and "people should live by the principle of obedience rather than reason." The theological packaging differs from tech's "trust the process" rhetoric, but the functional outcome is identical: dissent becomes morally impermissible. Open source BDFL structures typically lack any mechanism for removing a problematic leader except forking—the nuclear option of copying the codebase and starting a new project. As Eric Raymond theorized, the threat of forking theoretically keeps dictators benevolent. But the WordPress case demonstrates the flaw: when the BDFL controls the entire ecosystem infrastructure (domains, trademarks, update servers), forking becomes prohibitively expensive. The community discovered it had no brakes, no checks, no recourse except expensive litigation. Python's experience revealed another accountability gap: the unsustainable personal burden placed on a single individual. After the PEP 572 controversy, Guido van Rossum wrote: "I don't ever want to have to fight so hard for a PEP and find that so many people despise my decisions." He described feeling that opposition was "directed at him personally" and mentioned health issues "exacerbated by the continuous threat of always being the most-responsible person." NAR leaders similarly face either unchecked power or personal burnout—there's no structural middle ground. The accountability gap in both contexts produces predictable failure modes: - **No separation of powers**: One person controls spiritual/technical direction, financial resources, and community infrastructure - **No impeachment mechanism**: Leaders can only be removed by catastrophic exodus or legal action - **Bus factor vulnerability**: No succession planning because acknowledging mortality undermines divine/visionary authority claims - **Financial opacity**: The WordPress Foundation allegedly never disclosed to the IRS that it owns the WordPress trademark; NAR ministries like Sean Feucht's stopped publicly reporting finances after rapid growth --- ## Psychological dynamics that silence critics operate identically The most striking parallel between NAR and BDFL contexts lies in the psychological mechanisms used to manage dissent. Dr. Stephen Crosby, critiquing NAR authority doctrines, documented that "many who may ask an honest question of leadership are often accused of being disloyal, being 'wounded,' having a problem with submission to authority, or having a 'Jezebel spirit.'" The dismissive language differs in tech contexts—"not getting the vision," "not a team player," "bikeshedding"—but the function is identical: pathologizing legitimate concerns. **DARVO (Deny, Attack, Reverse Victim and Offender)** patterns appear consistently in both domains. Dr. Jennifer Freyd's peer-reviewed research demonstrates that DARVO is commonly used by individuals confronted about wrongdoing and is effective at manipulating third parties. Her 2020 study found that exposure to DARVO responses was associated with less belief of victims—but crucially, "DARVO-educated participants rated the perpetrator as less believable." Education about these dynamics can mitigate their effects. Institutional DARVO—when organizations rather than individuals deploy these tactics—appears in both contexts. When Bethel Church files copyright claims against critics, when Automattic's Matt Mullenweg frames himself as victim of WP Engine's "taking advantage" after his aggressive actions, when departing employees are characterized as "not aligned" rather than principled dissenters, the same pattern emerges: deny wrongdoing, attack the credibility of those raising concerns, reverse the moral positions of institution and critic. The silencing infrastructure extends to formal mechanisms. Christianity Today documented that "a growing number of ministers, missionaries, Christian workers, abuse victims, and victims' advocates are publicly objecting to the non-disclosure agreements and confidentiality clauses used by major religious organizations." The same pattern appears in tech: 159 Automattic employees who departed during the WP Engine crisis faced severance agreements; critics are banned from WordPress.org and company Slack channels. Danny Silk's _Culture of Honor_, used widely in NAR contexts, illustrates how "honor" language functions. While the principle of honoring people seems benign, critics note that "having a culture of honour often means **pulling your punches when you deliver legitimate criticism or not saying them at all.**" Tech parallels include "psychological safety" rhetoric deployed not to protect vulnerable contributors but to shield leadership from criticism. --- ## Case studies reveal the spectrum from burnout to corruption The BDFL case studies reveal a spectrum of outcomes—from structures that merely burned out well-intentioned leaders to those that enabled apparent corruption. **Python represents the burnout case.** Guido van Rossum was by most accounts a genuinely benevolent dictator who made good technical decisions. Yet the structure proved unsustainable even with an exemplary leader. The 18-month debate over PEP 572 (a relatively minor feature) exhausted him because "all opposition was directed at him personally." He stepped down not due to corruption but because the model required one person to absorb unlimited conflict. Python's successful transition to a steering council demonstrates that mature communities can move beyond BDFL governance when preparation, community maturity, and gradual power transfer are in place. **Linux represents the communication-harm case.** Linus Torvalds remained BDFL while acknowledging his abrasive communication style drove away contributors. Developer Sarah Sharp quit citing a "toxic community," asking whether "the community that built Linus up and protected his right to be verbally abusive will change." The 2018 adoption of a Code of Conduct—accompanied by Torvalds' acknowledgment that his "flippant attacks in emails have been both unprofessional and uncalled for"—represented structural adaptation without governance transition. The subsystem maintainer model had already distributed much operational authority even while Torvalds retained final decision power. **WordPress represents the conflict-of-interest case.** Unlike van Rossum or Torvalds, Mullenweg's situation involves direct commercial conflicts. His simultaneous control of Automattic (for-profit), WordPress Foundation (nonprofit), and WordPress.org (community) creates structural conditions where personal financial interests can drive community-harming decisions. The WP Engine lawsuit alleged "extortion, abuse of power, and tortious interference." A federal court granted WP Engine a preliminary injunction, finding Mullenweg's actions likely violated legal standards. This case demonstrates what happens when "benevolent" dictators have financial incentives misaligned with community welfare. **NAR scandals parallel the WordPress pattern.** Bob Hartley, a prominent NAR prophet, was eventually barred by both IHOP and Bethel Church after nearly a dozen women accused him of harassment and sexually exploitative behavior over a decade. The Todd Bentley Lakeland Revival saw elite NAR prophets sponsor and prophesy over a leader whose ministry collapsed amid personal scandals—yet no accountability followed for the failed prophecies. The Olive Heiligenthal case at Bethel, where a deceased child's resurrection was prophesied for days, illustrates how prophetic authority claims can persist without accountability for failure. --- ## The benevolent dictator problem transcends individual character The most important insight from this cross-domain analysis is that **good intentions do not prevent structural harm**. Alexandra Stein's research on cult dynamics found that "not all leaders want to get rich, gain sexual favours, or grab political power. But all want utter control over others." The desire for control itself—even when motivated by genuine belief in one's vision—creates conditions for harm. Organizational psychology research on "transformational leadership" reveals the mechanism. Dennis Tourish and Ashly Pinnington argue that these theories "have the potential to encourage authoritarian forms of organization... an overemphasis on the primacy of transformational leadership risks the promotion of undue conformity, the suppression of dissent, and the growth of cult-like dynamics." A key finding concerns **attribution bias**: studies of annual reports show that leaders attribute bad performance to external conditions while attributing good performance to their own decisions. This creates unfalsifiable leader infallibility—exactly the dynamic in both NAR ("God spoke through the apostle") and BDFL ("the founder's vision is why we succeeded") contexts. The research on organizational silence by Morrison and Milliken identified how "powerful forces in many organizations cause widespread withholding of information about potential problems or issues by employees." When combined with honor/loyalty cultures, DARVO dynamics, and NDA practices, even well-intentioned single-leader structures systematically suppress the information needed for course correction. Max Weber's concept of charismatic authority explains the pattern: such leaders "are not bound by tradition or rules and they may be accorded by their followers the right to pronounce on all aspects of their lives. The leader may lack any accountability, require unquestioning obedience, and encourage a dependency." Whether the charisma is framed as prophetic anointing or technical genius, the structural dynamics are identical. --- ## Cross-domain lessons illuminate solutions What can church leaders learn from open source governance failures? Python's successful transition offers a model: the PEP process provided infrastructure for discussing governance changes, the community was mature enough to handle democratic decision-making, and van Rossum participated in the new steering council initially to provide continuity while genuinely ceding power. The key design principle was that "the council should use powers as little as possible"—distributing rather than concentrating authority. What can tech communities learn from NAR critiques? The extensive theological literature on spiritual abuse provides precise vocabulary for dynamics that tech workers experience but struggle to name. Steven Lambert's "33 Signs of Spiritual Abuse" includes patterns directly applicable to tech: apotheosis of leadership, absolute authority, abuse of authority to coerce submission, constant indoctrination with "group" or "family" mentality. The Church of England's safeguarding manual now formally recognizes that "requiring unquestioning obedience—requiring obedience to the abuser, with an implicit or explicit suggestion that this equates to obedience to God [or the company mission]" constitutes abuse. Both domains point toward similar solutions: **The plurality model** addresses single-point-of-failure risks. Churches practicing plural eldership report that mutual oversight prevents abuse of power, shared wisdom improves decision-making, and distributed responsibility reduces burnout. Open source steering councils (Python) and Project Management Committees (Apache) serve similar functions. The key is that authority-holders must be genuinely accountable to each other, not merely advisory to a still-supreme leader. **Independence of oversight** is essential. As one church governance expert noted, "power to appoint = power to silence." When leaders select their own accountability boards, accountability becomes theater. Python's steering council cannot modify its own governing document without community vote; healthy churches separate board selection from pastoral control. **Ori Brafman's "starfish and spider" framework** suggests the optimal solution is hybrid rather than purely decentralized. Organizations need some centralized elements (core values, legal structure, infrastructure, tiebreaking authority) combined with decentralized operations (decisions close to the work, distributed leadership roles, community-driven selection, independent accountability). Pure BDFL fails; pure consensus also fails through paralysis. --- ## Building structures that survive their founders The path forward requires recognizing that governance structures must be designed to function even when leaders are less than benevolent—or simply human. Healthy accountability structures share common features across domains: **Clear decision-making processes** must be documented before crises. Python's PEP process existed before van Rossum stepped down; churches with clear bylaws fare better than those where the pastor's word is policy. **Term limits and rotation** prevent entrenchment. Python's steering council holds elections after each major release; Apache's PMC membership depends on ongoing contribution. Churches with term limits for elders develop leadership depth. **No-confidence mechanisms** provide escape valves. Python allows a two-thirds core team vote to remove steering council members; congregational governance traditions provide similar paths. WordPress's lack of any such mechanism left the community with only litigation as recourse. **Conflict of interest rules** matter critically. Python prohibits more than two steering council members from the same employer; healthy churches separate financial decisions from pastoral control. WordPress's crisis stemmed precisely from unaddressed conflicts. **Transparent financial management** prevents the pattern seen in both NAR ministries and foundation-captured open source projects. Regular audits, open accounting, and independent financial oversight are essential. The fundamental tension between efficient decision-making and healthy accountability is real but not unresolvable. Django's experience—multiple governance model changes as circumstances evolved—demonstrates that governance structures can adapt. The key is building in mechanisms for that adaptation rather than treating initial structures as permanent. ## Conclusion: Structure determines outcomes more than character The deepest lesson from comparing NAR governance failures with BDFL open source crises is that **structure determines outcomes more than individual character**. Van Rossum was by most accounts genuinely benevolent, yet the BDFL structure still burned him out. Many NAR apostles genuinely believe they're serving God, yet the structure still silences legitimate concerns and protects abusers. This insight should reshape how we evaluate leaders and organizations. The question isn't primarily "Is this leader good?" but "Does this structure have the accountability mechanisms to function when leaders are imperfect—as all humans inevitably are?" The absence of such mechanisms isn't neutral; it's a choice that systematically advantages power over truth. The WordPress crisis, Python's transition, Linux's Code of Conduct adoption, and ongoing NAR accountability failures all point toward the same conclusion: the BDFL model is, as Red Hat's analysis concluded, "mostly considered an anti-pattern." The question is whether communities will learn this lesson before crisis or after—and whether they'll build structures that can survive their founders' inevitable limitations.