Human-level Artificial Intelligence has now been developed and democratized. For the first time in human history, intelligence is digital and practically free. We believe this is an epochal moment as significant as the invention of electricity which transformed our planet over the last century.

Machines can now converse and solve problems but only the most enthusiastic early adopters take advantage of them. Much like electricity, we believe that general intelligence would only be truly revolutionary when it becomes easy to use and integrated into our lives and workplaces.

Many routine tasks require only minimal cognitive effort, e.g. filling forms, copying and pasting information from one tool to another, or making and sending reports. Human workers continue to be responsible for many tedious tasks that artificial intelligence is already capable of performing.

We believe this is because artificial intelligence is hard to use. It shouldn't require learning new tools. It shouldn't require context switching. It shouldn't require keeping up with the latest models and learning new prompting strategies, copying and pasting context, or wrangling with complex LLM APIs.

At General Agents, our mission is to liberate humanity from digital labor. We believe this happens by bringing artificial intelligence to where people already are.

This is why we are building Ace.

Ace leverages a new behavioral training paradigm. Unlike language and vision models which are trained on text and images, Ace is trained on behavior—the process that generates text, images, and other work outputs. Training on behavior generalizes better, as corroborated by the use of step-by-step reasoning in training frontier language models.

Creating behavior data is also more natural for domain experts, who simply need to record themselves performing a task using the tools they are already familiar with. They don't need to learn new tools or new processes. Ace is able to use the screen recordings, mouse and keyboard logs to learn how to perform similar tasks.

Because we're early in the scaling curve, Ace is not perfect. Ace is currently trained using a small fraction of the compute and data used by frontier language models. As we scale up, Ace will become more intelligent and capable.

We envision a world where Ace learns from the experience of every human and in return provides us its expertise and agency—empowering us to shape our lives on our own terms.