People & Agents
Software has two users now.
For decades, we’ve poured countless hours into building better software for people. Designing interfaces, mapping out complex workflows, obsessing over the smallest details and tweaking things over and over again to get them just right.
Then, seemingly overnight, everything changed. AI agents have become a huge part of how we work, writing code, managing tasks, translating emails, finding answers and taking on work that once meant navigating our way through half a dozen different applications.
It’s an exciting shift, but it raises an interesting question. What should software look like when the person using it isn’t always a person?
It’s time to simplify.
Software has spent decades getting bigger, accumulating more features, dashboards, settings and complexity along the way. We’ve reached a point where managing our software can feel like a job in itself.
We believe agents give us an opportunity to reverse that trend. Instead of building applications that try to do everything, we can focus on smaller, more useful tools and let agents take care of much of the heavy lifting.
A little less bloat and bullshit wouldn’t hurt either.
Agents are collaborators, not replacements.
We’re not building for a future where people disappear, far from it. We see agents as collaborators that enable us to do better work, make better decisions, explore possibilities that were previously out of reach and spend less time on repetitive tasks.
Most importantly, bringing agents into our teams doesn’t mean handing over our understanding of what’s happening. People still need to see what’s going on, explore and discover information for themselves, and make their own decisions. We still have a responsibility to think critically, refine ideas and bring the care and craft that makes good work great.
Build for both.
Until now, building great software has largely meant building great interfaces for people. That still matters, it really does, but we now have another user to consider.
The exciting part, at least from a design perspective, is that this gives us an opportunity to build even better interfaces for people. As agents take on more of the heavy lifting, we can move away from complicated workflows and endless configuration towards simpler, more focused ways to explore, understand and interact with our work.
Meanwhile, agents need their own way in. Access to information, context and the ability to take action without having to navigate the same interfaces we do. Neither should be an afterthought.
That’s what we’re building at Bake.
We’re building simple, focused tools for people and the agents they work with. Not software that tries to replace you, or another bloated platform with an AI chatbot bolted on, but useful things designed around a different way of working.
Because the future of software isn’t people or agents. It’s people and agents.