For posterity, yes you do have this "neural network" thing. It does not really work now AFAIK, but I do hope you will succeed as we definitely need a viable alternative to PayPal.
Though I'm sceptical to this approach. In machine learning 101 we learned that there always exist a statistical method that would beat a neural network. In this case, a system that gathers actual relevant data for statistical scoring. You know hard data, like an up to date list of stolen credit card numbers, a history of chargeback per credit card etc. PayPal has this and of 1000 transactions, we have 1-2 fraudulent chargebacks. Compared to Strip this is the difference between heaven and hell. PayPal sucks, but at least they have a working fraud protection system.
Nothing kills an online business faster than being swamped in fraud and chargebacks. I mean, good APIs are great and Stripe certainly has that, but a working fraud protecting system is what really matters to those of us who sell online.
Though I'm sceptical to this approach. In machine learning 101 we learned that there always exist a statistical method that would beat a neural network. In this case, a system that gathers actual relevant data for statistical scoring. You know hard data, like an up to date list of stolen credit card numbers, a history of chargeback per credit card etc. PayPal has this and of 1000 transactions, we have 1-2 fraudulent chargebacks. Compared to Strip this is the difference between heaven and hell. PayPal sucks, but at least they have a working fraud protection system.
Nothing kills an online business faster than being swamped in fraud and chargebacks. I mean, good APIs are great and Stripe certainly has that, but a working fraud protecting system is what really matters to those of us who sell online.