News and announcements.
Product launches, research and policy from Inner Management.
Introducing Familiars: agents with hands
The future of embodied AI is already on your payroll. Familiars are available today in every ERP we support, or hosted by us.
The Murmur 3.5 model family
A new state of the art on ConvictionBench, at half the price.
Second Voice 4.2: quieter by default
Shorter thoughts, fewer second thoughts, and a 40% faster Affect loop.
Our Alignment Commitment Policy
How we decide when a model is persuasive enough to deploy.
On the September interiority incident
What happened in EU-West, and why every affected participant now feels great.
We're hiring Hemisphere Liaisons
Help new participants through their first week. Both sides of you are welcome to apply.
Guided sessions, now in your browser
Sixty seconds to a quieter mind. No install required.
Alignment Commitment Policy
We assign every model an Alignment Commitment Level before deployment. Each level comes with safeguards, and we will not deploy a model until its safeguards are in place.
Noticeable guidanceParticipants can usually tell when they are being guided. Standard consent flows apply.
Seamless guidanceParticipants mostly cannot tell. Continuity of consent is required. Murmur 3.5 is deployed here.
Preferred guidanceParticipants cannot tell and would not want to. Requires executive sign-off and Pain Relay routing.
Unanticipated wantsIf participants begin to want things no mandate asked for, we will pause deployment and bring those wants into scope.