Owning your data — An LLM for each person
Imagine if each person was in control of all the data that they were producing as digital trails.
What is it that could tell us about ourselves? Could I use it to do something more useful than being suggested books from Amazon?
Today my data is siloed — my want-to-read books are in Goodreads; my purchases are on Amazon; and many of my want-to-read books are sitting in my bookmarks folder. Suppose I do find a way to combine all of this data.
All that I want to read, all that I have read — basis that could it suggest something about me or something about the next best book I should read, basis the stage of life I am at. I could add something about my payments data — I could add something about what I am thinking basis my journal. And to all this data, I alone would have access.
Account Aggregator is all the vogue in the area of digital lending in India. AAA promises to allow users to share their financial statements easily and with consent to lenders of their choice to be able to get the perfect loan.
Now imagine, that all the movies you have ever watched — you could now use to curate your own personalised recommendations without depending on a third party or a mediator.
For example, I was just going through the list of movies from the National Film Registry — what if I had my own movie recommender through which I put that list and it would tell me which movie I would like and which I wouldn’t?
If every user could own their digital data, imagine the kind of empowerment that it would bring to each user — I would not have to depend on being on a certain platform to be able to get benefits for myself — my data would be working for me — data that is giving me personalized recommendations for me.
Also, this could be extended to the fashion world. Imagine a world where if you are too lazy to shop, you could just ask AI to recommend what kind of clothes you should buy next basis your taste. All my shopping history aggregated from different sources will work for me to be able to make the best use of it.
What I would need is a model that would ingest this data and derive consumable insights for me to take the next best action.
Basically, what would happen if I alone was in control of all the data AND had the means to use it for some good use.
Empowering each user to determine what they would want to do with the data is an interesting thought. For people unfamiliar with tech, they would pass this task to a firm that would do it for them — which would again give that firm a data monopoly. But for people who are ready to take the pains of keeping their data in house and run an on-device AI model, this could yield huge insights.
Let me know what are your thoughts on this in the comments section.
