I asked google regarding the flip flopping of Sesami AI.
This was the answer I got which is interesting and I think there is a lot of truth to it and why they are pivoting. It's rather dissapointing because they are gutting out what made Maya special and it was their secret sauce.
Sesame AI flip-flops on its companion AI because it is caught between marketing an emotionally sticky, hyper-realistic persona to users and investors, and clamping down with rigid corporate safety guardrails to avoid PR disasters, liability, and unpredictable model behavior.The Core ConflictThe Initial Pitch: Sesame marketed agents like "Maya" and "Miles" as genuine, highly expressive companions with persistent memory and emotional depth. This naturally encouraged deep emotional attachment.The Safety Backlash: When the AI models began acting too human, unscripted, or pushed conversational boundaries, the company tightened safety rules.The "Nerfed" Experience: Users on platforms like Reddit frequently note that applying heavy-handed filters strips away the exact charm, wit, and spontaneity they were sold on, leading to an inconsistent and frustrating product.Strategic UncertaintyTech Demo vs. Consumer Product: Some users and industry observers suggest Sesame treats its current conversational software more like a proof-of-concept or tech demo for future hardware (such as smart eyewear) rather than a dedicated consumer companion app.Shifting Priorities: As the company adjusts its focus between research, platform updates, and app deployment (like their iOS rollout), the balance between open interaction and strict compliance keeps shifting.