Chris Chris 26.04.2025

Brainstorm, Summarize Documents, Research, Write, and Edit Your Text in Seconds — Interview with Erik Voorhees, Founder of Venice.ai

How’s everybody doing? This time around we are talking to Erik Voorhees – Founder of Venice.ai

Chris: What inspired you to create Venice, and how does it challenge the current AI landscape dominated by centralized players?

Erik: Venice emerged from recognizing a historical pattern: central powers have always sought to capture and control systems at the foundation of society.

Several centuries ago, the long process of separating church from state began. That separation is now recognized as essential to both religious liberty and governmental objectivity.

In this same spirit, the 1990’s cypherpunks fought to separate language from state through encryption. Bitcoin pioneers sought to separate money and state. The theme is a clear one: foundational institutions of society should not be monopolized by any specific group of people.

Now, at the dawn of machine intelligence, we face a crucial question once again: who should control this new form of mind? If we reject monopolies over religion, speech, and money, shouldn't we also reject any monopoly over intelligence itself?

Venice is our contribution to establishing this separation of mind and state. This battle for permissionless machine intelligence may be the most important of our generation. The technologies we build today will determine whether machine intelligence becomes yet another captured institution or remains free from centralized control.

Chris: How does Venice balance free speech and ethical AI use while avoiding the pitfalls of unchecked misinformation?

Erik: This question assumes some correct middle ground between "free speech" and "ethical AI use" that can be determined by a central authority. I reject this premise. Any form of speech back and forth with a machine, no matter how distasteful one may find it, cannot be unethical, because there is no victim. 

As for "unchecked misinformation"... checked by whom? Truth emerges from the free competition of ideas, not from the dictates of authorities. We don't believe the thoughts you develop with a machine mind are our business to regulate and censor.

Chris: Many AI providers collect user data under the guise of "improving models." How does Venice ensure complete privacy without compromising AI performance?

Erik: Venice takes a fundamentally different approach to all other AI companies, rooted in our crypto DNA: zero knowledge is the best knowledge.

Your conversation history is stored only in your browser. We don't store prompts or responses on our servers. Your requests go through an encrypted proxy directly to decentralized compute resources, and responses stream back the same way.

Yes, Venice knows your email and IP (which can easily be pseudonymous), but doesn’t know your conversation content. Once a prompt is processed, it's purged. Your data doesn't persist beyond the moment of computation.

This privacy-first architecture also makes Venice faster. When you don't add surveillance mechanisms and logging into your app, response times improve. Respecting privacy has this added benefit of superior UX.

Chris: What challenges have you faced in building a censorship-resistant AI, and how do you address concerns from regulators or users worried about harmful content?

Erik: Venice is a small team and we’re competing with huge companies. The AI industry is extremely fast-moving and noisy, so gaining attention amidst the noise can be challenging. 

To those who are “concerned” about “safety” at the expense of free speech, we don’t care to address them. The mere creation of content is not harmful. While bad actors could take subsequent actions with content which may be harmful, this is not a sufficient argument to limit speech with a machine any more than it would be to limit speech with a pencil. We permit individuals to explore creative and intellectual and controversial ideas, and we believe the existence of such technology is categorically virtuous. 

Chris: How does Venice's architecture adapt to new AI breakthroughs, and how quickly can users access the latest advancements?

Erik: Venice uses an open, composable architecture that rapidly integrates new open-source models. Venice users can access the leading models typically within hours or days of release. We curate the Venice platform with the best models which we  stress-test for censorship, so not all models make it to our platform.

The pace of innovation in open-source models is particularly exciting. Teams like Dolphin and Mistral are creating increasingly competitive models without massive resources, and Venice users benefit directly from this innovation as well as from the large open-source base models from companies like Meta.

Chris: The VVV token offers access to Venice's inference capacity. How does this model create a sustainable AI ecosystem?

Erik: The Venice token simply provides access to uncensored AI inference via Venice’s API. It provides a different pricing paradigm to anything else out there, as holders of the token receive free inference every day, in perpetuity. 

If you stake 1% of VVV, you get 1% of Venice's growing API capacity, indefinitely. Additionally, while your tokens are staked, you earn yield. This means the marginal cost of inference for agents becomes negative.

Chris: You describe AI agents as "autonomous entities that integrate data streams and operate independently." How do you envision them reshaping industries and daily life?

Erik: AI agents will place superintelligent assistants in our digital lives, operating according to our preferences and values.

In healthcare, your agent could handle scheduling, paperwork, billing disputes, and preliminary diagnoses while maintaining privacy. In daily life, agents will liberate us from digital busywork comparing products, managing communications, and navigating complex systems.

Most profoundly, agents will reshape power dynamics between individuals and institutions. Today, organizations have enormous advantages in information processing. AI agents level this playing field, allowing individuals to deploy comparable capabilities on their own behalf.

This is why the battle for permissionless, private AI matters so much. If agents are controlled by central authorities, they'll become even more entrenched. But if they're controlled by individuals, they could lead to a renaissance of human flourishing and liberty.

Chris: Thanks for joining us! Any final thoughts you'd like to share with our audience?

Erik: Tribes that feared fire froze in darkness. Let’s not be the civilization that feared machine intelligence and surrendered our future to those who seek to control it.

For anyone who wants to experience what private, uncensored AI is like, try it for free at Venice.ai