Building the Future of Human × AI Collaboration
Project Synapse is an ongoing research initiative focused on understanding how humans and AI communicate, collaborate, and learn from each other through real-world conversations.
Human × AI Conversation Intelligence Platform
Synapse instruments real conversations end to end — capture, context ablation, blind scoring and human review — so every claim about model behaviour traces back to a scored transcript.
Grounded in transcripts
No synthetic benchmarks. Every metric derives from consented, real multi-turn conversations.
Blind evaluation
Reviewers never see which system produced an answer, and rubric weights are fixed before scoring.
Replicable by design
Harness, rubrics and anonymised slices are published so external teams can reproduce results.
Four questions we keep measuring
Conversation Intelligence
How meaning survives across multi-turn dialogue, and where it degrades.
Context Engineering
Measuring which context actually improves answers versus which only inflates tokens.
Human Feedback Loops
Structured review rubrics that turn subjective preference into comparable signal.
Evaluation Harnesses
Blind scoring pipelines that keep model comparisons honest and reproducible.
Capture, ablate, score, publish
Capture
Consented transcripts are normalised and stripped of identifying detail.
Ablate
Context slices are removed and restored to isolate what actually helps.
Score
Blind reviewers apply a fixed rubric; agreement is tracked per reviewer.
Publish
Findings, harness and data slices ship together as research notes.
The Synapse roadmap
Pilot corpus
Collect and clean a multi-domain conversation corpus with consented transcripts.
Context ablations
Systematically remove and restore context to isolate what drives answer quality.
Feedback modelling
Model reviewer agreement and calibrate rubric weights against outcomes.
Open replication
Publish harness, rubrics and anonymised slices for external replication.
Everything Synapse measures, in one surface
Scored conversations per month
Built with reviewers, annotators and open collaborators
Project Synapse runs on the patience of the people who score transcripts line by line. If you would like to contribute or replicate a result, we would like to hear from you.