Experiments from Every’s Frontier team
What we tried this week at the edge of what is possible.
We test new models, agents and ways of working in our own work, every week. This is the log: what we tried, what we learned, and everything we wrote or recorded about it.
As the frontier advances, so do you.
Newest first
007
Toolbox
A place to share which AI tools you use.Exploring004Compounded knowledge recall
A test of whether an agent finds the right past learning when it needs it, and a CLI built to measure that.Exploring003Thinkroom
A place for living docs, where Kieran shares experiment reports for people to read and collaborate on.Exploring002Riffrec
Record feedback on a product and hand the recording to an agent. This round added live voice.ExploringEveryone sees each experiment, its cover screenshot and its one line. What we learned and the links open after a free Every sign-in.