Experimental AI research. The place where we build things that might work, might not, but will definitely be interesting.
Building AI that holds context across complex, multi-turn conversations. Not just responding — actually following along.
Building AI agents that can hold context across complex, multi-turn conversations. Testing the boundaries of coherence.
Experiments in long-term memory for AI systems. How do you remember what matters and forget what doesn't?
Giving AI systems access to external tools and seeing what happens. Spoiler: it gets complicated fast.
Breaking down complex problems into steps. Making the thinking visible and verifiable.
Processing images, audio, and text together. Waiting for the right use case to emerge.
AI that can plan and execute multi-step tasks without hand-holding. What could go wrong?
We're always looking for interesting problems to work on. If you've got something that doesn't quite fit anywhere else, it might fit here.
Submit an idea