We founded Imbue in 2021 to build an AGI future where humans remain at the helm, shaping powerful AI systems rather than being subordinated to them.
We believed that AI could multiply productivity and help everyone prosper. But as capabilities began to accelerate, a nagging worry set in: we felt forces dragging us all toward a future where powerful AI capabilities become concentrated in the hands of a few, giving them outsized control over what our chatbots say, what our AI agents do, and ultimately what institutions, societies, and lives we build. In this future, most people quietly become less free.
We’ve now come to understand something crucial: the core challenge of AI lies in managing how it shifts power.
AI makes software systems dramatically more powerful. That power flows, by default, to those who can build and own those systems. This concentrates power, which leads to exploitation and disempowerment. We’re starting to see it play out today: tech platforms use basic AI agents in the form of recommendation algorithms to hijack our attention, creating addictive experiences we simultaneously crave and resent. We have little control over how these algorithms operate — and they often don’t work in ways that benefit us.
At Imbue, we initially thought creating AI agents that helped people automate computer tasks would naturally distribute power. But over time, we began to see how this risked undermining people in a way similar to tech platforms: AI agent builders control the algorithms that make decisions about their users’ lives, and their incentives may not be aligned. Agents optimized for profit or engagement might nudge us toward buying sponsored products that advertisers want, gain access to trusted data because it’s lucrative to monetize, or manipulate our emotions to keep us engaged.
Instead of locking people into centrally-controlled agents, we had to rethink our approach: to genuinely put power back into people’s hands, we had to equip everyone to create, customize, and truly control their own AI tools.
Imbue’s mission is to empower humans in the age of AI by creating powerful computing tools controlled by individuals. We believe this requires a shift in the philosophy of building AI — not selling AI software designed to serve the interests of its creators, but instead helping us all make AI software that’s tailored to our own goals and values. For example, I’d love an agent that helps me track and participate in local ballot measures. Or a personalized feed that curates only the most important news to protect my attention from the latest flashy headline. Or an app that protects my grandmother from scam calls in Chinese.
Critically, this kind of software remains accountable to individuals and communities. Like the right to vote in a democracy, we believe the ability to create, modify, and control AI software and agents gives people a voice in this era of powerful AI — so that humans can be actors, not acted upon. In an era racing to make machines that replace humans, we want to reveal the builder in every human.
Today, we see glimpses of this possibility: AI coding tools seem tantalizingly close to letting anyone build software simply by describing it. But that initial spark of creativity burns out quickly when we try to really use these apps and we discover how flimsy, difficult-to-extend, and unmaintainable they are. We hit a ceiling because AI coding tools struggle to fix bugs without creating more of them, or to add new features without breaking old ones.
But humans know how to build complex software — we’ve spent decades developing best practices for architecting reliable systems, testing, and managing changes. Today’s LLM workflows rarely incorporate these practices. But what if they could? We’re trying to create a better way to build software with AI — one that embeds engineering best practices directly into AI-assisted software development, so more people can easily create robust, dependable software for themselves and others.
Our initial product is a coding agent environment that helps engineers write healthier code faster with LLMs by making it easy to encode best practices, identify and fix issues, and test and run LLM-generated code safely.
You can see more details and try it out here.
Ultimately, our goal isn’t just to help engineers, but to embed our collective knowledge of software craft into an open environment of coding agents that invites much broader participation. When we imbue engineering best practices into software creation tools, it becomes much easier for anyone to build sophisticated software, opening the door for many more people to participate in the AI future. Instead of waiting for companies to build for us, we’ll be able to make our own idiosyncratic tools for ourselves and our communities.
When we can create and control AI software and agents, power shifts. At the most basic level, being able to build our own interface to services lets us resist algorithms and platforms we currently cannot opt out of; I’d make my own Twitter feed optimized for thoughtful topics and friends, rather than one built to provoke me. When we can pivot to our own solutions, platforms are forced to better serve our interests to keep us engaged. The same dynamic applies when faced with other entities’ AI agents that try to influence us — for example, we can build our own filter agents that block spam or unwanted messages to safeguard our interests.
To help level the playing field, we also need laws and societal structures that let the agents we build be as powerful as those controlled by companies with lots of data (for example, by letting us get our own data out of corporate silos), and that protect us from other entities’ agents when they impinge on our freedoms (for example, by trying to addict us, warp what we believe to be true, or subtly shape our decision to buy something). This is the core of our policy work at Imbue: to safeguard individual rights in an increasingly automated world and uphold democratic principles against power concentration.
Technology’s highest purpose is not to replace human capability, but to amplify what is already inside us. We believe creative potential lies within every person, waiting to be unlocked. The world’s most meaningful software is still trapped in the human imagination, locked behind the barrier between what billions of people can imagine, and what they can actually create.
And if we free it, we can create something better than our current trajectory: a world more democratic, more open, more free. A world where we imbue machines with our will to shape our lives and institutions, where we collectively direct our agents toward solving what matters most to us.
Instead of AI replacing humans, we can use it to nurture what is best within each of us: the capacity for creation, connection, joy, beauty, awe.
This is the human future we can fight for. This is what AI ought to be for.