Arto Bendiken, CTO of Haltia.AI, on Privacy, Trust, and Ethical AI in the World of Personal AI

We thank Arto Bendiken, co-founder and CTO of Haltia.AI, who shares exclusive insights in an AI Time Journal interview. Dive into the world of privacy, ethical AI, and innovation as Bendiken discusses Haltia.AI’s unique approach, from on-device processing to the intersection of blockchain and AI. Learn about the transformative impact of Haltia.AI’s personal AI assistant on daily lives and businesses. Explore Bendiken’s autodidactic journey, leadership philosophy, and advice for tech entrepreneurs. This interview offers a visionary perspective on AI, privacy, and Haltia.AI’s pivotal role in reshaping the future of artificial intelligence.

With your background as a cypherpunk and advocate for privacy, how do you ensure that Haltia.AI adheres to these principles in its development and deployment?

At Haltia.AI, trust and trustworthiness form the core of our product ethos, particularly regarding security and privacy. For us, trust is not an abstract concept but a series of concrete actions that define our approach to personal AI. To begin, we enforce a strict policy where absolutely no user data is uploaded to the cloud. This isn’t just a standard procedure; it’s an ironclad rule that’s integral to our operations.

We’ve designed our system to respect user privacy inherently, meaning we don’t track our users or their interactions. Instead, we’ve devised what one might compare to the Windows Insiders program, where users can opt into Early Access for beta testing, providing anonymized metrics voluntarily. This select group likely represents 5-10% of our user base, but it’s from this data we can extrapolate vital business insights without compromising the privacy of the majority.

To safeguard privacy while still obtaining crucial business metrics—for instance, those critical for company evaluations and fundraising—we’re considering an opt-in program. Although it’s yet to be named, one could think of it as the ‘Haltia Insiders’ program, drawing a parallel to familiar models that users can immediately grasp.

Beyond these measures, our commitment to privacy extends to internal policies. Unlike some unfortunate instances in the tech industry where a staff member stole cryptocurrency from a customer server, our staff have zero access to user data, preempting any potential misuse. 

Moreover, to bolster trust further, we are initiating an independent security audit. This audit will scrutinize our data handling processes, the efficacy of our encryption both at rest and in transit, and attest that we adhere to the highest standards and to best practices. Our aim is not just to avoid negatives like backdoors, adware, or malware but to affirm our commitment to security proactively.

In conclusion, part of our advisory board comprises nine cypherpunks, reinforcing our dedication to privacy and security as foundational elements of Haltia. AI. Their expertise and oversight help us navigate the complex landscape of personal AI with a clear, unwavering commitment to the principles of the cypherpunk movement.

Can you describe a challenging technical problem that Haltia.AI encountered and how your team overcame it, particularly leveraging your ethos of ethical AI?

At Haltia.AI, we are still in the early stages of our journey, navigating the vast and complex terrain of artificial intelligence. As such, we have been fortunate to approach each step with careful deliberation, ensuring our ethos of ethical AI informs every decision we make. This forward-thinking approach has enabled us to preemptively address challenges before they manifest into technical problems. Our proactive stance means we are yet to encounter a technical hurdle that we haven’t been able to manage effectively with our current strategies and ethical framework.

As we progress, we expect to face and overcome challenges typical of innovative tech startups. When we do, our commitment to ethical AI will be the lens through which we find solutions, ensuring that every product development, every feature enhancement, and every aspect of user interaction is not only technically sound but also ethically responsible.

As someone who has been at the forefront of blockchain innovation, how do you see blockchain technology intersecting with AI to enhance privacy and security?

The intersection of blockchain and AI is a fascinating domain, particularly when viewed through the lens of enhancing privacy and security. Illia Polosukhin, my former boss at NEAR Protocol and co-author at Google of the seminal paper ‘Attention Is All You Need,’ which sparked the current AI renaissance, is pioneering efforts to marry these two technologies. At Haltia.AI, we’re approaching this nexus with a cypherpunk’s vigilance for trust and privacy.

Blockchain excels at establishing consensus among mutually untrusted parties, enabling what’s termed ‘trustless’ interactions. ‘Trustless’ here doesn’t imply that the system is untrustworthy. On the contrary, it means that the system operates in a way that does not require users to trust one another or a third party — the integrity of transactions or interactions is maintained by the blockchain’s protocols and cryptography.

This trustless coordination can serve as a robust mechanism for Haltia.AI members, ensuring that operations such as messaging could be encrypted on the blockchain, ensuring privacy and security by design.

In the open-source AI space, there are no guarantees against data misuse — your data could be leaked, inadvertently used to train algorithms, or misused by a staff member. At Haltia.AI, we’re building something fundamentally different. Our commitment to ethical AI means erecting ironclad guarantees around user data. It’s not just about building trust; it’s about enshrining it in our architecture.

So when we talk about Haltia.AI’s approach, it’s deeply rooted in trust. By potentially integrating blockchain’s trustless systems, we aim to provide a double layer of assurance — not only does our on-device AI eschew the need for cloud-based data harvesting, but we also envision a future where every user interaction is verifiably secure and private, backed by blockchain’s immutable ledger. This dual approach, leveraging blockchain as a coordination mechanism for our members, could be revolutionary, ensuring that our personal AI operates within a framework where trust is implicit and security is a given.

Given your rich history as an autodidact in the tech field, what advice would you give to aspiring tech entrepreneurs who are just beginning their journey?

Embracing the role of an autodidact in tech has instilled in me a deep appreciation for a foundational focus. My advice to new entrepreneurs? Resist the urge to build upon the unstable ground laid by others, the strata of half-arsed implementations compounding to vast incidental complexity. Instead, devote your energy to mastering the core principles, the essential complexity. When you understand the fundamentals intimately, you’ll often find yourself questioning the paths others take, revealing that they, perhaps, have strayed from these essential building blocks.

In the world of technology, where evolution is relentless, this foundational focus becomes your compass. It equips you to navigate through noise and distraction, steering you toward innovation that is both grounded and original.

Adopt a healthy skepticism towards established norms.

Arto Bendiken

Furthermore, adopt a healthy skepticism towards established norms. The challenge received wisdom. In technology, the status quo often represents solutions that are ripe for reevaluation and improvement. If something appears to be a half-measured approach, it likely is, and therein lies your opportunity to innovate. So, question everything, trust in your foundational knowledge, and be ready to pave a new path that others haven’t dared to tread. This is the essence of tech entrepreneurship — it’s about being bold enough to build the future you believe should exist. There is no shortage of low-hanging fruit for someone with a fundamental focus to pick.

Editor note: see also “7 Cardinal Mistakes Startups Make with VCs: Insightful Lessons from the Front Lines By Talal Thabet, CEO and Co-founder of Haltia.AI”

How does your philosophical viewpoint influence the leadership and strategic direction of Haltia.AI?

At Haltia.AI, our philosophical stance is deeply rooted in the open exchange of ideas, which starkly contrasts with the prevalent IP regime that thrives on exclusivity. We’ve consciously chosen not to engage in the patenting race — a game we see as costly and detrimental to innovation in the industry. Our approach is grounded in the belief that progress is hampered by imaginary property monopolies, often referred to as IP.

This belief shapes our strategy in a significant way. Rather than hoard our innovations behind patent walls, we opt for open publication as a form of self-defense. If we originate an idea and share it publicly, it becomes part of the global conversation — a prior art that cannot be later weaponized against us or the industry. This is critical for us, particularly as we anticipate the field will undergo consolidation through mergers and acquisitions, where sizable IP portfolios will be used aggressively.

By engaging openly in the marketplace of ideas, we provide timestamped proof of our innovations, which not only fosters a collaborative environment but also protects our developments from becoming a weaponized commodity in the future.

Moreover, our competitive edge is not hinged on patents but on our ability to act swiftly and execute effectively. We recognize that any competitive advantage is transient; what matters is execution: the velocity and precision with which we bring our advancements to market. In this vein, our leadership and strategic direction are about fostering a culture of agility and execution. If we cannot maintain the pace required to leverage our innovations, then, frankly, we don’t deserve to lead. We thrive on our own merit, relying on our capacity to innovate rapidly and ethically without resorting to government-created monopolies.

In essence, Haltia.AI’s philosophy champions a competitive landscape where competence and execution prevail, and where the true ‘weapon’ is not patents, but the ability to innovate transparently and collaboratively, safeguarding the ethos of ethical AI and cypherpunk values we uphold.

Could you share insights on how Haltia.AI’s AI Assistant is pioneering in terms of on-device processing and privacy?

At Haltia.AI, we are certainly at the forefront of exploring on-device AI processing. While there are a select few venturing into this space, we are pioneers, particularly because we don’t just settle for what’s achievable now; we’re actively pushing the boundaries of what’s possible through rigorous research. We are channeling significant efforts into advancing the performance of AI processing within the device itself. This work involves not only our own internal research team but also collaborations with universities. These partnerships help us to delve deeper into the potential of on-device AI and to stretch the current capabilities further than what’s been done before.

It’s ironic yet enlightening to recognize that our contemporary ‘pioneering’ efforts to keep all user data exclusively on the device mirrors the norm from just a decade and a half ago—before we were beguiled by the mirage of convenience offered by cloud computing, leading us to relinquish vast amounts of personal data. The true cost to privacy has since become apparent, breaches are routine, and the potential for misuse, particularly by insiders at cloud companies, is a persistent threat—likely more prevalent than we know or is reported.

This move away from cloud storage, a pivot back to the foundational principles of data privacy, aligns with our ethos of not just building trust but embedding it into the very fabric of our technology.  By championing on-device processing, we’re defining a strict separation of domains: what is processed locally on the device and what, if anything, interacts with the internet. This stance is as pioneering as it is ethical—it’s about reviving an era where privacy wasn’t an afterthought.

In technical terms, the key breakthrough that has enabled this shift is quantization techniques. These techniques have made it feasible to host large-scale AI models on personal devices without compromising performance. 

This focus on privacy, combined with our cutting-edge on-device processing capabilities, sets Haltia.AI apart as a pioneer in the AI industry, not just in theory but through tangible, applied innovation.

As AI continues to advance rapidly, what do you believe are the key ethical considerations companies should prioritize?

As AI progresses at a breakneck pace, it’s crucial for companies to engage with ethical considerations fundamentally. Currently, there seems to be a lack of consistency in how ethical concerns are addressed, particularly in tech hubs like Silicon Valley, where the perspective can be quite insular.

The first step I advocate for is transcending our own narrow viewpoints: “Forgive [the barbarian], for he believes that the customs of his tribe are the laws of nature!”. We must aim for universal ethical principles that transcend local or tribal biases. This is vital because adopting a narrow ethical view based on the idiosyncrasies of a particular culture or region won’t translate globally and will inevitably lead to conflict and division. The polarized state of American politics is a testament to what happens when societies become too entrenched in their own narratives.

Taking a leaf from international commerce, successful global operations have shown us that imposing our ethical views is not only impractical but counterproductive. In a business context, and by extension in AI development, it’s not prudent to be overly concerned with the personal beliefs or behaviors of those you engage with, especially when such concerns do not affect the core operations.

The key, then, is to distill ethical considerations to what anthropologists term ‘human universals’—the core values shared across all cultures and societies documented in all times and places. These include fundamental principles like the prohibition of murder, which can be safely integrated into AI systems as ethical standards. Delving into more contentious areas risks alienating parts of your audience and is best approached with caution.

My aim here is to illustrate the complexity of ethical considerations in AI and the importance of adopting a universal approach that resonates with humanity as a whole.

What has been your approach to building a “dream team” of engineers, and how do you foster innovation within your team?

At Haltia.AI, our philosophy centers on fostering ‘extreme ownership’ among our engineering team. This concept shapes our approach to building what we consider a dream team. We grant our team members and managers a high degree of autonomy, with the understanding that autonomy is intrinsically linked to personal accountability. When challenges arise, we look inward, recognizing that the responsibility for both successes and setbacks starts and ends with us individually.

We have little patience for a culture of finger-pointing. We are cultivating an environment where each team member is not only encouraged but expected to take initiative, act proactively, and own their projects fully. This ethos is diametrically opposed to a disengaged ‘I just work here’ mindset that I detest. At Haltia.AI, ‘I just work here’ does not exist; our team is composed of individuals who know where the ‘bathroom’ is, metaphorically speaking and are ready to lead the way.

We also believe in rewarding this sense of ownership, not just through verbal recognition but through tangible ownership incentives that align personal success with that of the company. By eschewing micromanagement in favor of mentorship and support, we nurture rapid personal growth and development. This approach has consistently unlocked the creative potential within our team, leading to innovative breakthroughs and a dynamic, problem-solving culture.

How do you envision the personal AI assistant from Haltia.AI changing the daily lives of individuals and the operations of businesses?

At Haltia.AI, we fundamentally believe that the most precious commodity in our modern life is time. Our personal AI isn’t just a product; it’s a catalyst for transformation — it’s about offering what money can’t buy. We are in the business of giving time back to people and businesses, enabling them to win at life and excel in the market.

For individuals, this means a personal AI that streamlines daily tasks, manages schedules and processes information with unprecedented efficiency. It’s about creating more space for creativity, relaxation, and life’s joys by offloading the cognitive load of routine tasks to an intelligent companion. Your adaptive AI evolves with you, learning from your habits and preferences to serve not just as a tool, but as a long-term partner in navigating the complexities of day-to-day life.

For businesses, our AI will redefine operations by automating workflows, enhancing decision-making with real-time data analysis, and providing insights that drive efficiency and innovation. It’s about enabling teams to focus on strategic initiatives and creative solutions rather than getting bogged down in administrative details. By doing so, businesses can operate more dynamically, adapt faster, and maintain a competitive edge.

In both cases, it’s about winning — winning back time for what truly matters, whether it’s personal growth, family, innovation, or customer engagement. Haltia.AI’s vision is to empower people and organizations to operate at their highest potential, making every second count.

Considering the competitive landscape of AI technologies, what unique capabilities or features does Haltia.AI offer that set it apart from others in the field?

In the frenetic advancement of AI technologies, Haltia.AI stands out with a foundational approach that veers from the status quo. In a landscape bustling with activity but scant on substance, many companies amount to little more than façades built upon the infrastructures of giants like OpenAI. They ride the waves of open AIs, yet this reliance has shown its frailties — from outages to cannibalization of features by the very platforms they depended on, a risky proposition for any business’s longevity and a painful lesson for investors.

At Haltia.AI, we diverge from this precarious path. We don’t merely harness existing tech; we innovate from the ground up, ensuring that our core capabilities are proprietary and deeply integrated — except for the foundational model Mistral 7, where all our technology is developed in-house, a testament to our commitment to deep tech.

Moreover, amidst the recent tumult at entities like OpenAI, where even colossal funding does not immunize against user experience woes, Haltia.AI prioritizes an exquisite user experience. Our approach mirrors Apple’s ethos of vertical integration, refusing to cede control over critical aspects of our business and the user experience, ensuring every element aligns with our standards of quality and reflects our core values.

Thus, what sets Haltia.AI apart is not just our unique technological advancements, to be disclosed at a later stage, but also our unwavering commitment to privacy, security, and a user experience that’s not just functional but delightful. We envision a product that embodies trustworthiness, where every interaction is secure, private, and precisely tailored to the user — or as we prefer to say, member. This comprehensive, member-first philosophy is what defines Haltia.AI in the competitive landscape of AI technologies.

What are some of the top books that have significantly helped you in your career, and what key takeaways from them have you applied in your executive roles?

In the vast sea of literature that has shaped my professional ethos, pinpointing favorites is as challenging as it is reductive. However, ‘Extreme Ownership’ by Jocko Willink and Leif Babin has been pivotal, explicating a leadership model that I followed prior to knowing its name, and which resonates deeply with Haltia.AI’s culture. This book is not just recommended but is part of the assigned reading for every manager in the company, reinforcing the notion that accountability and leadership are inseparable.

Alongside this, ‘Radical Candor’ by Kim Scott and ‘The Five Dysfunctions of a Team’ by Patrick Lencioni stand as crucial texts in our management training. These works offer profound insights into the subtleties of team dynamics and the art of honest communication, principles we hold paramount in our executive roles. The former is also interesting in highlighting and critiquing euphemistic management culture in America, while the latter explains why attempts to build remote-only teams so often fail.

While these titles form the bedrock of our management philosophy, they are but a sliver of the array of books that inform the philosophical underpinnings of our company’s direction. From tactical management to the broader philosophical discourse, the spectrum of reading that informs our approach is broad and deeply ingrained in our strategic thinking.

Who are some of the individuals who have significantly influenced your leadership approach or whom you follow for fresh perspectives?

My leadership approach has been profoundly shaped by a host of thinkers and doers. At the forefront is Paul Graham, the founder of Y Combinator, whose writings have been a staple of my reading for the last fifteen to twenty years. His insights into startup culture and innovation have offered invaluable guidance.

The ethos of Y Combinator permeates my perspective, so it’s not just Paul but also luminaries like Sam Altman, including prior to his tenure at OpenAI, who have influenced my thinking on running innovative ventures.

Jocko Willink’s work on ‘Extreme Ownership’ has provided a blueprint for leadership that resonates deeply within Haltia.AI, advocating for accountability and decisiveness, and I’ve learned from all his books, lectures, and so on.

Additionally, the input from my time working with Illia Polosukhin and Max Zavershynskyi at NEAR Protocol has been instrumental. Their experiences, such as Max’s insights drawn from his time at Google, have given me a vantage point on effective company management and team dynamics.

These are some of the influences that have contributed to a leadership style that emphasizes autonomy, ownership, and a relentless pursuit of excellence.

About The Author

Scroll to Top
Share via
Copy link
Powered by Social Snap