Digital Legacy Avatars

Most people don't realize how much is lost when someone is gone. Not just memories. The way they explain things. The way they tell stories. The tone, the pauses, the small details that never get written down. We store photos and videos, but they are fragments. They don't answer back.

The idea here is simple. What if memory could be interactive?

Instead of passively storing moments, we let people record themselves in a structured way. Stories, advice, opinions, context. Not random videos, but guided conversations. Prompts that pull out how they think, not just what they did. Over time, this becomes a dataset of a person, not just content.

That dataset can then be turned into something usable. An avatar that can respond. Something a child or grandchild can ask questions to. Not in a sci-fi sense, but in a grounded, useful way. "What would you have done?" "What was your childhood like?" "How did you think about money?" The goal is not perfection. It's familiarity.

Most products in this space get two things wrong. They either focus too much on the technology, building something impressive but emotionally empty. Or they lean too far into sentiment, creating something that feels uncomfortable or forced. This needs to sit in between. Technically simple, emotionally right.

The hard part is not the model. It's the input. If the data is shallow, the output will be shallow. That means designing the right structure, the right prompts, and the right experience for people who are not technical and often not comfortable on camera. This is more product design than AI.

There are also real constraints. Consent matters. Boundaries matter. This should never feel like replacing someone. It should feel like preserving access to how they thought and spoke. Something families opt into with clarity, not something imposed.

If done right, this becomes a new category. Not storage. Not memorials. A layer between generations.

We're looking for someone who understands that this is not a novelty. It's a product that sits at the intersection of memory, identity, and time. Someone who cares about getting the experience right, not just the demo.

What we want

• Input-first thinking: solve capture, not just generation
• Taste and emotional correctness over model complexity
• Strong structure: prompts that extract how people think
• Simple, fast UX that real families will actually use
• Narrow, opinionated wedge with real users early

What we don't want

• GPT wrapper with a UI
• Deepfake-first, tech-over-product thinking
• Anything that feels morbid, uncanny, or like replacement
• Overbuilt, time-consuming workflows
• Broad "AI memories" positioning with shallow execution