This is an interesting post and the person achieved a huge amount.
They should not have given up.
I would have pivoted to personal pet companions. If people want virtual girlfriends then they certainly want personal dogs, cats and birds that talk to you.
Enough was achieved here to raise money.
Some careful redesign and development might have allowed some caching and reuse of expensive AI API interactions, reducing costs.
Add in further gameplay elements so it is not just about making the AI API do stuff, that would also reduce costs and maybe even engage the users more.
Perhaps AI should be the icing and accents on a game rather than the focus.
As an aside, this is the first time I have read anyone tell about how they wanked to their own software.
Altogether this was perhaps a missed opportunity for this founder.
BUT it is now an opportunity for one of the HN readers to follow this playbook and replicate what this founder did.