You nodded along to the whole episode. So did we. We're building Bias Scanner to listen again, a bit slower, and show you where a good-sounding argument quietly wasn't one.
Launching February 2027
Two options are offered as the only ones, when middle paths exist (staggered days, trials by team). Then one step is said to lead inevitably to the next, with no reason given for why it would.
Scores the reasoning in this episode, not the people in it.
A meeting recording, a YouTube video (where captions are available) or a podcast. Upload the file or paste the link.
We plan for it to transcribe the audio, work out who's speaking, find the claims and look at how each one is argued.
Jump to each moment, hear it again, and read why it's shaky and what a fairer version would say. We plan for each flag to link to its illustrated Homo Rationalis card. Then decide for yourself.
Did we decide on the evidence, or did the loudest voice win? With everyone's consent first, and never for performance reviews.
Lectures, interviews and classroom clips, where captions are available.
A map of the episode's weak spots, so you can go straight to the few minutes that matter.
We plan to flag a move in an argument, never the person making it. No scores for people, hosts or sides.
We plan for every flag to carry the exact quote, a confidence level and a button to disagree.
We plan to test with mirrored arguments from opposite views. If only one gets flagged, that's our bug.
Meetings only with everyone's consent, and never for performance reviews. We plan to delete files automatically after analysis and to use only providers that don't train on your content.
Launching February 2027
One email when Bias Scanner launches. Unsubscribe any time. Privacy
We'll email you once, in February.