The End of the Assistant Era
Every AI conversation starts from zero. That's not a bug — it's a design philosophy that's already obsolete. Here's what comes after the assistant.
The Humanized AI · est. 2025
Intelligence becomes useful when it understands the person in front of it.
We study how people think, feel, and decide.
Then we build AI that actually reflects that.
What we believe
Most AI is built on a wrong model of what a human being is.
— Our founding observation
The implicit assumption everywhere in the field: we are rational agents who occasionally feel things. Emotion is treated as noise — a deviation from the clean signal of preference and intent.
The reality runs the other way. We are emotional beings who happen to reason. Every decision, every memory, every relationship is colored by affect before logic ever touches it.
When you build intelligence on top of the wrong model, you get something that is technically impressive and subtly wrong — and you can feel the wrongness without being able to name it.
Read the full manifesto 01 Product live
Nodymic →
Nodymic →
05 Founding principles
shape every decision
shape every decision
03 Ways to work together
See services →
See services →
Live 2025 →
Nodymic
The mirror for what you create
Emotional content analytics for brands. Understand how your audience actually feels when they encounter your content — before you publish it.
Visit nodymic.com
In development 2026 →
[Personal companion]
The mirror for who you are
An AI that truly knows you over time. Not a chatbot — a system that builds a model of you: your patterns, your blind spots, your becoming.
Coming soon
01
Understanding before optimization
You cannot serve someone you do not understand.
02
The emotional layer is not noise
It is the signal. Cognition runs on it.
03
A model of you, not of users
Demographics are not understanding. You are.
04
Serve, don't extract
Engagement is not the same as flourishing.
05
Humanized AI is more accurate AI
Modeling humans as they are predicts better.
Work with us
Building something that needs to understand people?
Technology licensing · Co-development · Humanization audits
AI Pulse · Weekly Brief
One finding a week that changes how you think about AI.
Cognitive science, neuroscience, and human-AI research — curated for builders. Not news. Implications.
You're in.
First brief lands in your inbox this week.
human-ai-interaction
Apple's $250M Siri Settlement Puts a Dollar Figure on Expectation Mismatch
Apple reached a $250M settlement over misrepresentation of Siri's capabilities, with the legal record indicating users formed expectation-based relationships with the assistant — grounded in anthropomorphic marketing framing — that the product could not sustain. Users did not respond to underperformance with rational recalibration; the documented response pattern was affect-laden betrayal, consistent with violation of a perceived relational contract rather than disappointment at a tool's limi…
ai-companions
Social Context Modulates Every Lifestyle Factor's Effect on Brain Health
A large-scale analysis of UK Biobank data, published in Nature Communications, found that brain health outcomes are determined by the interaction of lifestyle, environmental, and social factors — not their independent additive effects. The same behavioral inputs produced different cognitive outcomes depending on social context; isolation significantly amplified cognitive decline associated with negative lifestyle factors, while social connection attenuated it. AI systems that function as soci…
cognition
Removing Learning Friction Improves Session Metrics, Degrades Skill Formation
Papers within a Frontiers in Psychology research collection on generative AI and cognitive mechanisms examined the 'desirable difficulty' principle — the established finding that effortful processing improves long-term retention and skill transfer — and found that AI tools eliminating cognitive friction produce measurable short-term performance gains while masking long-term skill degradation. The effect on recall and transfer tasks is only visible in assessments conducted well after the learn…