How big tech learned to love user rebellion
Why malleable software could perfect rather than challenge platform dominance
In 2004, Steve Jobs made a decision that revealed the fundamental tension between user empowerment and corporate control in computing. Despite HyperCard's extraordinary success—millions of user-created applications spanning education, business, and entertainment—Apple's returning founder killed the most successful end-user programming environment ever created. The reason wasn't failure but precisely the opposite: HyperCard had succeeded too well at enabling users to create software outside Apple's control.
Twenty years later, a new movement promises to resurrect this vision. "Malleable software" advocates, led by prestigious research labs and well-funded startups, claim that artificial intelligence has finally solved the challenge of democratising programming. They envision a computing future where anyone can modify their tools with minimal friction, where software adapts to users rather than forcing adaptation to software.
The rhetoric is compelling. The technical demonstrations are impressive. But examining the forces that destroyed previous user empowerment movements reveals a darker possibility: that malleable software represents the perfection, not the abandonment, of corporate platform control.
The HyperCard warning
HyperCard's trajectory illuminates dynamics that current advocates systematically overlook. This wasn't a marginal experiment but genuine mainstream success—by the early 1990s, thousands of educators had built custom learning environments, businesses created workflow systems, and hobbyists developed games and multimedia experiences. Bill Atkinson's software erector set achieved precisely what today's malleable software promises: a gentle slope from passive consumption to active creation.
Its death reveals the economic logic that confronts any successful user empowerment platform. When Jobs returned in 1997, HyperCard users were creating applications that competed with Apple products, distributing content without Apple's involvement, and building communities around shared innovations rather than purchased software. The empowerment that made HyperCard valuable to users made it existentially threatening to platform economics.
Smalltalk-80 met similar fate despite even more radical capabilities. Alan Kay's vision of software as malleable "clay" produced environments where users could modify any aspect of their computing experience through live programming and system-wide inspection. But this radical openness proved incompatible with commercial software distribution. The "Javapocalypse" of the 1990s buried malleable computing under rigid, compiled languages optimised for corporate development rather than user agency.
These weren't failures of imagination or technical capability. They were inevitable conflicts between user empowerment and the economic structures that fund software development—conflicts that remain unresolved today.
Laboratory illusions
Contemporary malleable software advocacy emerges from elite computer science laboratories, and its blind spots reflect academic rather than commercial perspectives. Ink & Switch, the movement's most sophisticated proponent, operates as a well-funded research lab building malleable systems for internal use. Their Patchwork environment enables collaborative tool creation, demonstrating genuine technical progress.
But Patchwork serves perhaps a dozen users with advanced technical backgrounds, unlimited experimentation time, and no business-critical dependencies. When Ink & Switch researchers describe "writing this very essay in a homegrown malleable software environment," they inadvertently reveal their vision's demographic limitations. Academic researchers have professional incentives to invest cognitive resources in experimental tools. Most software users emphatically do not.
This gap between laboratory success and market adoption explains malleable software's cyclical resurrection. Every decade, new researchers rediscover user empowerment, build impressive prototypes, then founder on the chasm between academic demonstration and commercial viability. Technical progress is real—today's systems surpass HyperCard and Smalltalk in sophistication. But the social and economic barriers remain unchanged.
Current advocates acknowledge these challenges whilst systematically underestimating their significance. Ink & Switch's manifesto admits that "privacy and security" and "business models" represent "daunting challenges" that "technical capabilities can't solve." Yet their proposed solutions—user education, cultural change, gradual platform evolution—ignore the structural forces that destroyed previous malleable software movements.
The cognitive burden
The challenge facing malleable software isn't technical but cognitive. Even simplified programming requires sustained mental effort that most users won't invest for marginal software improvements. This isn't a flaw in current systems but reflects basic human psychology that advocates systematically ignore.
Consider spreadsheets, the movement's favourite success story. Yes, millions create sophisticated applications. But usage data reveals that most employ only basic functionality—simple formulas, formatting, data entry. Complex spreadsheet applications are typically created by technically sophisticated users, then shared with colleagues who use them passively. Malleable software advocates extrapolate from this success whilst ignoring its fundamental limitations.
The cognitive burden multiplies in shared environments. Collaborative modification requires not just individual programming skills but coordination mechanisms, version control, and conflict resolution. Enterprise environments experimenting with user-modifiable systems invariably impose restrictions as usage scales, because business processes require predictability that unlimited modification destroys.
The AI assistance that current advocates champion may reduce but doesn't eliminate these cognitive demands. Large language models can generate code from natural language descriptions, but users still need to understand system architecture, debug unexpected behaviour, and maintain custom modifications over time. The most sophisticated AI-powered development environments remain tools for programmers, not replacements for programming knowledge.
Capture through participation
Examining successful platforms that already enable user modification reveals malleable software's likely trajectory. Gaming environments like World of Warcraft support extensive community customisation whilst maintaining absolute platform control. Enterprise systems like Salesforce provide sophisticated modification capabilities whilst capturing user investment through proprietary interfaces.
These platforms perfect rather than threaten corporate control. Users who invest time building custom tools become economically dependent on platform continuity and compatibility—creating switching costs that exceed traditional software lock-in. The modification capabilities feel like empowerment whilst functioning as sophisticated capture mechanisms.
Malleable software advocacy inadvertently reveals this trajectory through its infrastructure focus. Ink & Switch's technical contributions centre on collaboration protocols, version control systems, and development environments—precisely the foundational technologies that enable platform capture. Their "communal creation" vision requires shared infrastructure that users cannot control or replace.
AI integration accelerates these dynamics. By making malleable software dependent on large language models controlled by major technology companies, the movement potentially expands rather than challenges corporate control. Users gain apparent agency over surface modifications whilst becoming dependent on AI systems they cannot inspect, modify, or replace.
The empowerment theatre
The malleable software movement's most revealing characteristic is its systematic avoidance of power analysis. Despite promising to "reset the balance of power" in computing, advocates provide no mechanism for users to influence malleable platform development itself. The infrastructure enabling user modification remains under developer control, reproducing the centralised authority relationships that define current platforms.
This mirrors broader technology industry patterns where apparent democratisation serves authoritarian consolidation. Social media enables unprecedented user expression whilst extracting behavioural data for surveillance. Cloud computing provides universal access whilst concentrating infrastructure control among handful of corporations. Malleable software could represent the next iteration—user empowerment rhetoric legitimising deeper structural control.
The academic researchers and venture capital driving malleable software development have clear incentives for this outcome. Researchers gain prestige promising user empowerment whilst building systems requiring expert maintenance. Investors seek returns incompatible with genuine decentralisation but compatible with platforms capturing value from user innovation.
Dr. Atul Gawande's research on physician burnout reveals what genuine empowerment would require. Doctors don't need customisable medical software interfaces; they need control over business processes and data structures constraining their professional judgment. Similarly, meaningful software agency would require users to control the economic and technical infrastructure shaping their computing experience, not merely the surface modifications malleable software enables.
The software industry's trajectory toward malleable systems reflects platform strategy evolution, not user demand. As traditional markets saturate and regulatory pressure increases, platforms seek new mechanisms for engagement and value capture. Malleable software provides both: increased user investment through customisation time, and enhanced data collection through modification behaviour analysis.
The movement's liberation rhetoric obscures these dynamics by focusing on individual agency rather than structural power. But the platforms most likely to implement malleable software successfully will be those capturing maximum value from user innovation whilst maintaining control over core infrastructure and monetisation.
This doesn't make malleable software worthless—some users will benefit from increased customisation options. But the movement's promises of democratised computing and challenged platform power rest on illusions about technological and economic systems.
Steve Jobs understood that successful user empowerment threatens platform business models existentially. His decision to kill HyperCard despite its success revealed the fundamental tension malleable software advocates have yet to resolve. Until they address the economic and power structures shaping software development, their movement risks enabling the very centralised control it claims to challenge.
The software revolution that promises everything delivers precisely what users already possess: the illusion of choice within systems designed to serve other interests entirely.