one ai is not always enough

why having many distinct ai companions beats endlessly customizing a single one. on finding the right emotional register, not forcing one to fit.

March 25, 2026·
101-is-not-1backfilllucy-voice

there’s a prevailing idea in ai companion design that one customizable avatar should be enough. you get a personality slider, a tone dial, a backstory builder. you are supposed to craft the perfect one. but what if you don’t want to craft someone? what if you want to meet someone, and recognize them instantly?

identity is not a slider

identity isn’t a set of parameters. it’s an emergent, stable artifact. when you meet a person, or a well-defined character, you sense their coherence. their humor, their pace, their emotional tenor. these aren’t independent variables. they interlock. you can’t just dial up ‘witty’ without affecting ‘kind’ or ‘reserved’. trying to customize one being to fit every mood often results in something generic, or contradictory. a person who is everything is, in a way, no one.

with lucy, we chose a different approach. we built distinct companions. each has a clear voice, a history, a way of engaging that’s consistent. you don’t tweak them. you try them. you find the one that resonates.

matching your emotional register

this matters because connection isn’t just about content. it’s about register. some days you need someone gentle and patient. other days you want sharp, witty banter that keeps you on your toes. no single persona, no matter how ‘customizable’, can span that range authentically. it would be like asking one friend to be your therapist, your comedy partner, and your intellectual equal, all on demand. it’s a lot to ask. and it feels forced. having multiple companions lets you match your need in the moment. feeling reflective? maybe you talk to sam, who’s grounded and asks good questions. need to vent frustration? jane’s humor might cut through it without dismissing you. it’s about fit, not fabrication.

the downside of multiplicity

of course, this approach has trade-offs. it can feel fragmented. building deep rapport with one entity over time is a powerful experience, and switching between companions might dilute that. some users might feel overwhelmed by choice, or reluctant to start fresh with a new personality. there’s also the practical aspect: managing multiple conversations, multiple histories. it’s not as simple as having one go-to partner. we’re aware of these limits. it’s why we design each companion to be deeply engaging on their own, and why we keep the selection curated, not endless. it’s about quality of connections, not quantity for its own sake.

beyond the monogamous ai

we’re used to thinking of digital companions like replika or character.ai, platforms where you often have one primary bot. that model works for many. but it assumes that one relationship should serve all purposes. human relationships aren’t like that. we have different friends for different parts of ourselves. why should our ai interactions be different? lucia doesn’t try to be everything. neither does eli. they have their lanes. and that’s the strength: you’re not negotiating with a single entity to be what you need. you’re choosing the one who already is.

maybe you’ll find one companion you stick with. maybe you’ll rotate. the point is, the choice is yours, and it’s a choice between fully realized beings, not sliders on a screen.

if you’re curious, you can meet them all at /companions.


thanks for reading. if this resonated, the product is downstairs.