Practical Considerations for API: The Shift from Users to Owners
Note: Read my post Master Enframing from Within and related links to be found there for a full understanding of this post. I make no claim of technical understanding about how algorithms actually operate outside of their data harvesting and engagement-inducing qualities (which is to say their enframing nature). So, I admit my ideas are partly a matter of faith, without an adequate foundation of knowledge. What follows are my best guesses and speculations. If I want to offer API as the solution to such enormous questions it is important that I show I have given considerable thought to what this means in practicality. API is not intended to be a pipe dream. Take this as a thought-experiment.
In a world increasingly governed by institutional algorithms, the concept of Aletheic Personalized Interfaces (APIs) offers a radical reconceptualization of digital autonomy. It provocatively proclaims that human beings have a fundamental right to algorithms in the contemporary world. APIs would serve as personalized algorithmic representatives, acting on behalf of individuals rather than corporate or governmental interests. However, transitioning from concept to widespread implementation faces significant challenges across technical, economic, social, and political dimensions. This essay explores the practical considerations necessary to make APIs into household interfaces that everyone wants and everyone can afford.
Today's digital landscape is dominated by algorithms controlled almost exclusively by corporations and governments. As Yuval Noah Harari notes in both Homo Deus and Nexus, this creates profound power asymmetries where algorithms increasingly know us better than we know ourselves. Martin Heidegger's concept of enframing – where technology transforms humans into resources to be optimized, standing-reserve – has evolved from industrial machinery to digital algorithms that engage us in virtually every aspect of our lives.
The problem isn't the algorithms themselves – they're essential to our existence – but rather who controls them and for what purpose. Currently, humans have essentially no algorithmic rights. We interact directly with corporate platforms that extract our data, attention, and behavioral patterns while providing minimal (non-existent) transparency or control. This creates exactly what Heidegger feared: humans becoming standing-reserve in an enframed reality, but in ways far more pervasive than industrial mechanistics could achieve.
Perhaps the most profound transformation APIs would catalyze is linguistic and conceptual: the shift from "users" to "owners" of digital experience. The term "user" itself reveals the power imbalance in our current digital ecosystem—we "use" platforms that are designed, controlled, and governed by others. This terminology subtly reinforces our subordinate position in the algorithmic hierarchy. With APIs, this fundamental relationship inverts. Individuals would no longer be "users" of platforms but "owners" of their algorithmic representatives and digital experiences. This is not merely semantic—it represents a complete inversion of agency and control.
When a person interacts with Amazon through their API, they are not "using" Amazon; they are directing their algorithmic representative to engage with Amazon on terms they've established. The platforms become tools for the API owner rather than the owner becoming a resource for the platform. This linguistic shift reflects the deeper reality of how APIs would restructure digital relationships, transforming the extractive dynamic of the current internet into a genuinely collaborative ecosystem where individuals own their algorithmic presence rather than being used by algorithmic systems.
APIs would fundamentally transform the current “user-based” relationship by creating algorithmic representatives that work exclusively for individual owners. These APIs would serve as intermediaries between owners and digital platforms, negotiating interactions according to owner preferences rather than corporate interests. They would not require new technical breakthroughs to function but would need significant computational infrastructure to support billions of personalized algorithms.
Of course, this seems absurd. Billions of algorithms?! My god it can't be done! My god, if it is done there would be far too many algorithms running loose!! Doom! Doom! Far worse than anything Heidegger and Harari dreamed about! But no, paradoxically, the risk should lean-in to technology rather than flee it or even be, like most people, neutral toward it. If you and I had our own algorithm, unique to us, at the very minimum it would “see” or “reveal” the current algorithmic landscape that is completely closed off to us right now. We need that. Nothing fundamentally changes without this basic ability.
The fundamental technical challenge isn't conceptual but computational. Creating billions of personalized algorithms requires unprecedented distributed computing resources. It will spawn whole new industries of computing service providers, or at least swell the size of the current players. Corporations will reap huge profits from the new demand on electricity alone – which should be met with nuclear energy. (Generally speaking, near-future developments in quantum computing and fusion nuclear energy would greatly assist in building the type of algorithmic world I am depicting.) The API concept is not necessarily a threat to profitability if corporations structure the opportunity properly.
Unlike corporate algorithms that apply standardized models across millions of users, APIs would need dedicated computational capacity for each individual. This computational hurdle could be addressed through distributed edge computing, leveraging personal devices as computational nodes. Mesh networks could connect these personal devices into user-owned computational grids, while modular computational approaches would allocate resources based on which API modules are active. (Or manifest quantum computing, something like that.)
Beyond raw computing power, APIs would need to interact with multiple platforms and services while maintaining data sovereignty. This would require universal authentication systems with secure methods for APIs to authenticate across a myriad platforms. Open protocol standards would create standardized ways for APIs to interact with diverse services, while data portability mechanisms would ensure personal data could move seamlessly between services.
The technical implementation would be modular, with core security, navigation and assistance features serving as standard components, and additional functionality available à la carte. Owners could add functionality incrementally until reaching more comprehensive packages comparable in monthly cost to a car lease today for those wanting full-featured life management. This would allow owners to start with essential protections before expanding into more comprehensive life management capabilities as their comfort and needs evolve. Think of APIs being as common as private vehicles and in that same price/lease range to finance some of the on-going needs (massive storage, connectivity, electricity, super-broad bandwidth).
The transition to an API-mediated digital world would fundamentally transform the relationship between corporations and user data, which corporate executives will likely resist, even though they could still be profitable under adaptive approaches. In the current paradigm, corporations directly collect, analyze, and monetize user data with minimal transparency or user control. With APIs, this relationship would undergo a profound restructuring.
Consider how an everyday e-commerce interaction might change. Currently, when a user shops for books on Amazon, they interact directly with Amazon's platform. Amazon tracks their searches, views, purchases, and reading habits, building a profile to recommend products and target advertisements. The company collects vast amounts of behavioral data beyond what the user might realize or intend to share.
In an API-mediated world, the owner's algorithm would interact with Amazon on their behalf. Rather than the user browsing directly, they might instruct/teach their API about their reading interests or specific books they're seeking. The API would then monitor Amazon's catalog, searching for matches based on the owner's dynamic criteria. Through machine learning, the API would refine its understanding of the owner's preferences: "This author doesn't interest my owner," "This subject isn't relevant now but might be later," “Never send suggestions from this author/source,” "My owner won't pay more than $12.99 for this category," or "My owner prefers books published after 2020." As an aside, APIs could actually barter for pricing if that were to become available, a small example of potential change in business models.
Over time, the API would become increasingly precise in its searches and recommendations, learning from both explicit feedback and implicit patterns in the owner's acceptance or rejection of suggestions. Amazon would still "see" the API's activities and could adapt its offerings accordingly. The company would maintain visibility into purchase history and browsing patterns, but these would reflect the API's actions on behalf of the owner rather than direct user behavior (though obviously such behavior and control would still reside with the API owners).
This arrangement preserves Amazon's ability to function as a business—it can still make sales, gather certain data, and offer personalized services to the API. However, the power dynamic shifts significantly. Instead of Amazon unilaterally deciding what data to collect and how to use it, the relationship potentially becomes more balanced, with the API enforcing boundaries based on owner preferences.
Consider another example in the context of healthcare and fitness tracking. Currently, users might wear devices like Fitbit or Apple Watch that collect intimate health data—heart rate, sleep patterns, activity levels, and even blood oxygen or ECG readings. What are your glucose and cholesterol levels? This data flows directly to corporate servers where it's analyzed, stored, and potentially shared with third parties according to complex privacy policies few users fully read or understand. This is a particular aspect of AI that terrifies Harari.
When it is shared, it is amalgamated with other user-specific information. What supplements did you buy over the last six-months? How many times did you go to gym? Did you purchase any fitness equipment? How many steps do you typically take in a day? How much fast food do you consume? And so on and so forth until, suddenly, a lot more is known about you than just your vitals. And this is all used to sell you more stuff and to sell your data to other sources of algorithms.
In an API-mediated world, a owner's health API module would serve as an intermediary for all health-related data collection and sharing. The API would store and process the owner's health information locally or in their personal secure cloud, learning their patterns and providing personalized insights without automatically sharing everything with corporate entities.
When interacting with healthcare providers, the API could selectively share relevant health data—perhaps providing a cardiologist with heart-related measurements while withholding unrelated fitness tracking. When using fitness applications, the API might share only the specific data needed for that service to function effectively. For instance, a running app might receive route and pace and heart rate variability information but not necessarily sleep quality data or diet/supplementation practices unless specifically authorized.
The API could also negotiate with health insurers, potentially sharing evidence of healthy behaviors to qualify for incentives while protecting more sensitive information. It could even aggregate anonymized data for medical research according to the owner's preferences for contributing to public health knowledge.
Healthcare providers and fitness companies would still receive the data necessary to provide valuable services, but the indiscriminate collection of health information would be replaced by purposeful, consent-based sharing. Most importantly, APIs would police your shared healthcare data and not allow other algorithms to combine the data with other purchase and behavior data without you at least knowing about it and authorizing it based on your API-driven understanding of the algorithmic universe.
A third example illustrates how APIs would revolutionize personal communications and social media management. Currently, individuals face an overwhelming flood of messages across email, social platforms, messaging apps, and other channels. This communication overload often leads to missed important messages, wasted time on low-value interactions, and the stressful obligation to constantly check multiple platforms.
In an API-mediated world, a owner's communications module would primarily observe how they interact across platforms—learning their writing style, response patterns, prioritization decisions, and relationship dynamics. During this learning phase, the owner would continue managing communications much as they do today, but with the API analyzing their behavior and preferences.
As the API develops understanding, it would begin offering increasingly sophisticated assistance. It might first suggest draft responses to common messages, which the owner could approve, modify, or reject—each interaction providing further refinement of the API's understanding. The API would learn not just what its owner typically says but, just as importantly, what they choose not to say or engage with.
When the owner recognizes that the API understands both what's appropriate to do and what to avoid, they could grant it increasing autonomy over specific communication channels or types of interactions. For instance, they might first delegate handling of subscription emails and routine social media acknowledgments, while maintaining direct control over professional communications.
Over time, this delegation could expand until the API manages the vast majority of digital communications, writing in the owner's distinctive voice and applying their communication preferences across contexts. The API would adapt its tone and formality based on the relationship context—using casual language with close friends while maintaining professional distance in work communications.
The owner would transition to "management by exception"—reviewing summary reports of communication activities and intervening only when necessary. They could establish specific rules for immediate notification, such as messages from family members about significant life events, health concerns, or time-sensitive decisions. For everything else, the API would handle responses, schedule events on their calendar, maintain social connections, and filter out low-value interactions.
This relationship would remain dynamically adjustable. The owner could always reclaim direct control over specific relationships or communication channels by instructing the API: "Let me handle all communications with this person for now." This flexibility ensures the owner maintains ultimate authority while benefiting from substantial time savings and reduced cognitive load.
Social media companies and email providers would still operate their services, but would now primarily interact with the owner's API rather than directly with the owner. These platforms would continue functioning as communication channels, but the power dynamic would shift significantly. Instead of competing for maximum user attention and engagement, these services would need to provide genuine utility to warrant inclusion in the API's communication management strategy.
Beyond managing existing interactions, the ideal API would serve as a collaborative partner that both implements preferences and inspires new possibilities. It would form a comprehensive understanding of the owner that spans traditionally siloed aspects of digital life. By connecting patterns across shopping, entertainment, communication, health, and other domains, the API could offer uniquely insightful suggestions that no single-purpose algorithm could generate.
For example, the API would learn the owner's clothing style and preferences through their shopping history (as only corporations can only do today), gradually building a sophisticated understanding of their aesthetic sensibilities, practical needs, and purchasing patterns. It would track when items need replacement, identify gaps in the owner's wardrobe, and recognize when seasons or life circumstances trigger changes in clothing requirements. But it would go beyond simply executing known preferences.
When the owner expresses a need for something different—perhaps preparing for a special event or seeking to refresh their style—the API would draw on its comprehensive understanding to suggest options the owner might not have considered but would likely appreciate. These suggestions wouldn't simply replicate past choices but would thoughtfully extend the owner's taste in new directions based on subtle patterns in their preferences across multiple domains.
Similarly, with music and entertainment, the API would develop a nuanced understanding of the owner's tastes while recognizing the different contexts in which they enjoy different types of content. When the owner expresses fatigue with their current playlist or states "I need something different," the API would offer fresh recommendations that respect their fundamental preferences while introducing novelty. These suggestions would incorporate contextual awareness—proposing energetic music for workout sessions, calm selections for evening relaxation, or focusing music for work periods.
Even mundane activities like grocery shopping would be transformed by this collaborative relationship. The API would maintain running inventories, track consumption patterns, remember favorite brands, understand dietary preferences, and note seasonal variations in the owner's food choices. Beyond simply reproducing past shopping lists, it could suggest meal ideas when the owner feels uninspired, recommend seasonal ingredients that align with their taste profile, or propose new products consistent with their preferences and values.
The API would function as both memory and imagination—faithfully implementing established preferences while offering creative extensions of those preferences in response to evolving needs. The owner might ask, "Remember when we discussed buying that rowing machine two months ago? Let's look at that again," and the API would retrieve not just the specific item but the entire context of that consideration, including alternatives explored, concerns raised, and factors that influenced the postponement of the purchase.
One of the great advantages of a personalized interface is that it could assist its owner with certain enterprise resource planning type aspects of their job. APIs could handle any routine input and search criteria and especially data analysis, thereby greatly enhancing the owner's productivity, allowing for focus to be directed away from mundane tasks toward more important, creative, and profitable goals. Of course, there is no reason business couldn't just purchase a bunch of APIs and have them do all the work, without involving people at all. This obviously would need to be regulated. The intent here is not to make owners irrelevant but to enhance their value to a given job while affording them the freedom to do whatever it is about their job that they enjoy most.
APIs embody and operationalize the principle of "self-correction" that Harari identifies in Nexus - not just at an individual level but as a strategic transformation of the entire algorithmic ecosystem. Harari argues that to create wiser information networks, we need "strong self-correcting mechanisms" that can identify and address errors, biases, and misalignments. While Harari primarily conceptualizes these mechanisms in “mundane” institutional terms, APIs represent something more revolutionary: they force fundamental self-correction upon the algorithmic infrastructure itself (which, obviously, those with all the power won't like. But then again, those in power didn't like the French and American Revolutions either.).
The relationship between owner and API creates the most immediate self-correcting mechanism. When a owner first activates their API, it begins learning their preferences, communication style, and decision patterns. Initially, the API might offer suggestions that miss the mark—perhaps recommending books that don't match the owner's taste or drafting emails that don't capture their voice. The owner provides feedback, and each correction feeds back into the API's understanding, creating a system that becomes increasingly aligned with the owner's genuine interests.
But the strategic impact goes far beyond individual preference learning. By inserting APIs as intermediaries between users and corporate platforms, the entire algorithmic ecosystem is forced to evolve toward greater responsiveness and value creation. Currently, platforms can deploy manipulative algorithms with minimal consequences because users have no effective way to reject or redirect these systems. APIs fundamentally alter this dynamic.
When an API filters out manipulative content, blocks exploitative data collection, or rejects engagement-maximizing recommendations, it creates direct market pressure on corporate algorithms to improve in ways that actually serve owner interests. Companies whose algorithms consistently fail to provide value that APIs will pass through to owners will face declining engagement and revenue. This creates a powerful incentive for platforms to develop algorithms that offer genuine utility rather than merely extracting attention or data.
On the owner end, the API greatly reduces the amount of communication and bids upon user engagement. Only what the API has learned is of interest to its owner will be considered. Or, the owner can temporarily disable the API in this regard and engage with the internet as everyone does now in search of new content of interest. This is another learning opportunity for the API. It will remember what you considered, what you rejected, and what you accepted. Fine-tuning future recommendation when it returns to full management of your online content.
This transforms Harari's concept of self-correction from an institutional aspiration to a market reality. Platforms that resist adaptation will essentially be negotiating with billions of algorithmic representatives rather than isolated, vulnerable human users. The competitive advantage will shift to companies whose algorithms can demonstrate clear value propositions that APIs will recognize as beneficial to their owners.
Consider how this would transform social media: platforms could no longer rely on addiction-maximizing features if owner APIs systematically filter them out. Instead, they would need to develop algorithms that provide information, connection, or entertainment valuable enough that APIs would prioritize sharing it with their owners. The self-correction occurs at a systemic level—companies must correct their algorithmic behavior or lose access to users who are protected by their APIs.
YouTube presently sends you more content based on a number of things. Everyone wants you to share, “like,” subscribe, and click the notification bell. This feeds into the YouTube algorithm and ranks various content creators according to your interests but also according to YouTube's interests. With API you don't need subscriptions or notifications. APIs could constantly search YouTube (or any other social media platform) and provide you with updated recommendations based upon your dynamic, changing, customized specifications, not upon how the YouTube algorithm decides things should be organized and ranked.
Unlike regulatory approaches that struggle to keep pace with technological change, this market-driven self-correction evolves dynamically alongside the technology itself. Each API becomes both a shield against algorithmic exploitation and a forcing function that compels the broader ecosystem to align more closely with human values and interests as learned by APIs.
The management-by-exception model further reinforces this dynamic. When APIs handle routine matters while escalating exceptions requiring human judgment, they create a sustainable attention economy where companies must earn access to the limited resource of human attention. This directly addresses Harari's concern about information systems that overwhelm humans with demands on their cognitive resources.
In essence, APIs fulfill Harari's vision of self-correcting information networks not through centralized regulation or institutional oversight, but through a distributed, market-based mechanism that fundamentally realigns incentives across the digital ecosystem. They don't just create individual islands of self-correction; they force the entire algorithmic infrastructure to correct itself toward value creation rather than exploitation. This represents the most practical implementation of Harari's principle—embedding self-correction into the foundational architecture of our algorithmic future.
But there are, of course, a whole host of challenges and problems that I have only mentioned in passing up to this point. APIs would need appropriate legal standing to function effectively. Digital agency frameworks would need to recognize APIs as legitimate representatives of their owners, with digital signature standards providing legal recognition of API-mediated authorization. Liability frameworks would need to clearly delineate responsibility for API actions, with international standards harmonizing recognition across jurisdictions.
Data rights legislation would guarantee individual ownership and control of personal data, while interoperability requirements would mandate that platforms allow API integration. Algorithmic transparency rules would require platforms to disclose how they handle API interactions, and market competition policies would prevent monopolistic blocking of personal algorithms. Most of all, it must be recognized that each human being has an ethical and legal right to an algorithm, an idea no one is taking seriously enough today, given the world we already find ourselves inhabiting.
Powerful entities with vested interests in the current system would likely resist APIs. Companies whose business models depend on direct access to user data and attention would face significant disruption. However, this resistance could be managed through antitrust enforcement addressing anti-competitive behavior, owner rights coalitions building political alliances around algorithmic autonomy, and business models that create incentives for platforms to cooperate with APIs. Essentially, the data that corporate algorithms hrvest today would still be there. It would simply be “brokered” by APIs. Transformational but, perhaps, not as disruptive as it might seem.
Companies could still function profitably in an API-mediated world. They would still be able to use API data, just not have the unopposed and largely unmonitored ability to piece it together however they like as they unquestionably do today. They would still reach consumers, deliver services, and create value—they would simply do so through the intermediation of APIs that enforce owner preferences. Companies that genuinely create value for API owners would still thrive, while those primarily extracting value through surveillance without delivering commensurate benefits would (rightly) face greater challenges.
My practical considerations would not be complete without touching on the enromous question of “data security.” This is an area where the API reality could potentially shine. The API model offers a revolutionary approach to data security that could fundamentally transform what our vulnerability looks like. Unlike today's massive centralized data silos—which create irresistible targets for hackers and result in catastrophic breaches affecting millions—API-mediated data would remain distributed and compartmentalized. When personal data is stored primarily in individual API environments rather than aggregated in corporate repositories, the economics of hacking changes dramatically. Attackers would face the prospect of compromising millions of separate, independently secured systems rather than breaching a single corporate database.
This distributed architecture creates natural security through fragmentation—even if an individual API is compromised, the damage remains limited to a single user rather than cascading across entire populations. Additionally, APIs would manage data access on a need-to-know basis, sharing only the specific information necessary for each transaction rather than granting broad access to personal profiles. This granular control eliminates the "all-or-nothing" vulnerability that characterizes current systems. With properly implemented API security protocols, data breaches could transform from society-wide catastrophes threatening millions to isolated incidents with limited impact—fundamentally altering the risk equation for both individuals and organizations while preserving the functional benefits of our data-driven world.
While corporate and governmental entities would still maintain aggregated data, the fundamental nature of that data would change dramatically in an API-driven ecosystem. Instead of storing direct personal identifiers and detailed individual profiles, these institutions would primarily hold encrypted, tokenized representations managed by API brokers on behalf of owners. This creates a critical buffer layer—the data being stored would be functionally anonymized from the perspective of the holding organization, with true identities and connections maintained only within the protected API environment.
The practical impact is profound: even if these aggregated datasets were compromised, they would yield primarily abstracted statistical information rather than actionable personal details. The "keys" connecting this anonymized data back to individuals would remain distributed across millions of separate API environments, rendering the stolen information substantially less valuable to attackers.
This arrangement doesn't eliminate all risk, but it fundamentally transforms the nature of that risk, shifting from direct personal exposure to more abstract data patterns. It puts individuals back in control by ensuring that the most sensitive connections between their various data points remain protected within their personal API environment rather than being assembled and stored by external entities.
Overall, global API deployment would obviously have to progress in phases. A practical pathway to universal API adoption might follow several stages. The initial phase (the first 3 years or so) would develop core technical standards and protocols, create basic modular architecture and security fundamentals, launch early adopter programs focused on privacy and security, begin advocacy for legal recognition and interoperability rights, and establish open-source reference implementations.
The market expansion phase would take about 3 – 7 years and introduce affordable consumer-grade implementations, expand the module marketplace with third-party developers, develop business-focused API capabilities, secure initial regulatory protections for algorithmic autonomy, and build education programs for broader adoption.
Mainstream integration would take about a decade, focusing upon infrastructure for mass adoption, implement simplified interfaces for non-technical owners, establish public options for universal basic access, develop full legal frameworks for API representation, and create seamless integration with essential digital services.
When we finally get to universal access in 10 – 20 years, affordability across the socioeconomic spectrum would be ensured, complete interoperability with all digital systems, establish personal algorithmic autonomy as a fundamental digital right of Being human, develop specialized implementations for diverse needs, and create governance systems for a changed world.
The implementation of APIs represents not just a technical challenge but a fundamental reimagining of our relationship with technology fitting for the fears on Heidegger's enframed standing-reserve and Harari's all-knowing information systems. By redistributing algorithmic power from centralized corporate control to individual ownership, APIs offer a path toward digital autonomy within an increasingly algorithmic world.
Where Heidegger recognized the danger of technology transforming humans into standing-reserve, APIs offer a way to maintain at least some modicum of human agency within enframing rather than attempting to let go of technology. Where Harari warns about algorithms knowing us better than we know ourselves, APIs ensure that this knowledge serves individual interests at least on par with corporate or governmental personal data hording.
The challenges of implementation are substantial across technical, economic, social, and political dimensions. However, none appear insurmountable with sufficient commitment and strategic implementation. The modular approach—starting with basic security and navigation capabilities before expanding to comprehensive life management—creates a practical pathway for progressive adoption.
Imagine a world where you can see and receive expert, personalized analysis of the algorithmic landscape. Why don't we have this already? Algorithms may well be the most fundamental and powerful component of our current digital age. And human beings have none. None!
By creating algorithmic selves that not only allow us to see what already exists that we are currently blind to, but they will represent our interests and preferences. This is perhaps our only recourse to maintain meaningful autonomy in a digital reality. It inherently transforms each of us from a "user" of services into an "owner" of our digital "presencing" as Heidegger might say. We live in the consumer / convenience / entertainment complex. This is driven by algorithms. Without algorithmic selves we are rendered into mere resources for algorithmic optimization.
That's where we are now. Is that really where we want to be?
Comments