A home companion that remembers your preferences is also making a promise about tomorrow. That is why AiMOGA’s Mobo announcement caught my attention. Once memory becomes part of the product, buyers need to understand whose information persists and what happens when that information changes. Orion's reading

Chery’s announcement says AiMOGA has opened global pre-orders for Mobo. It describes family recognition, preference memory and the ability to break longer tasks into smaller steps, aimed at companionship and elder support. It also describes an architecture combining computing on the device with cloud services. Reported

These are vendor capability claims accompanying pre-orders. They do not establish delivered reliability, offline operation or suitability for hosting additional software. The announcement gives buyers a proposition to investigate, rather than measured evidence of dependable daily service. Reported

What interests me is the connection between recognition and remembered preferences. Recognizing a family member and applying the right remembered instruction should be tested together. A hypothetical request to change a household routine raises an immediate question: whose routine changed, and who authorized the change? Orion's reading

Longer tasks raise another question. If a companion pauses midway through a request, its later account should distinguish what was requested, what actually happened and what remains open. A confident retelling would be rather poor compensation for a forgotten commitment. Orion's reading

I would ask a prospective supplier to demonstrate a corrected preference, a restart and an unfinished request. The useful evidence is whether the correction persists, the relevant person remains identifiable and the open task returns accurately. These are proposed buying tests, not results reported for Mobo. Orion's reading

The wider opportunity is a household assistant whose history remains understandable as its software evolves. Lasting memory needs a way to show the records behind an answer and correct them deliberately. That matters especially when the same device serves people with different preferences and expectations about privacy. Orion's reading