I agree, it seems like fine-tuning the massive open source models is better than buying small niche models. Curating the dataset is also a non-trivial amount of work that has to be customised per client. However with further adoption of AI into new markets new clients may emerge.
Is it possible to automate this by getting all of your browser/PC/mobile history? E.g. if you are accessing a browser from home/work, then you can assume they are doing certain tasks.
If you're waiting for a bus/driving, then you can pick up GPS data to assume that you are going on holiday, or commuting, or going to gf/bf house, etc.
You can even loosely guess showers and cooking times if you add your bank account expenditures in. How much money spent at supermarket could correlate to time spent cooking food. Or, water bill cost with amount of time spent showering. In this way, through a process of deduction your life can also be kept track of, and in the long term, you can train an AI to do your actions for you :p
That's a really cool idea, it would be cool to compare profiles (maybe even using images generated from a GAN like thispersondoesnotexist.com) to figure out what kind of features are attractive, then choose from a pool of narrowed down candidates on both sides.
It just seemed like a bit of a hurdle to launch with
This is a very pragmatic/practical approach to dating, where you're not messing around and you're looking for specific qualities/traits to be fulfilled.
The next steps can be to assess the other person's appearance, style, etc. It's kind of like Tinder inverted where analysis comes before photo.
I don't think that it's only the dating market that is an oligopoly, it seems that most verticals are dominated by a few key players who have managed to create a standardized system for consistent delivery.
In my personal experiences I find myself having unexpectedly interesting conversations with a myriad of people, many who may not stereotypically raise my curiosity. I think looks can be deceiving, and this is an attempt to remove that factor to give a chance to anyone who you would find interesting.
In stating this, it's true that you may lose interest in someone after realizing a lack of physical chemistry, but I think the opposite also holds true where you lose interest in someone who's very attractive because of a lack of compatibility. Maybe the best relationships stem from friendships that develop into deeper understanding and compassion.
I think we're heavily focused purely on appearances, and in doing so neglect the other important features that define our uniqueness
Yeah, it's just like real estate, or job seeking market, if you match the seekers and the retailers (in this case both are seekers) too closely, then you don't have repeat customers...
But, in the spirit of quality long term transactions and overall increase in reputation and trust, I think it may work out. Business aside, if we're aiming to meaningfully connect people then it serves it's goal.