This is generally true but not always. As you can imagine we have loads of data on this. If you only selected students who had been tinkering with writing code

This is generally true but not always. As you can imagine we have loads of data on this. If you only selected students who had been tinkering with writing code for 10 years certainly you'd be successful in doing so, but you'd also eliminate ~90% of those who we have seen become software engineers. The only way we've found that does it well is to have people actually start writing code and see if they enjoy it. That's why we now have multiple free classes, have a free dropout period once you're in the school, and even have a three-week free trial of the school itself. The notion that financing ISAs removed the incentives isn't really accurate. First, most of the time ISAs are financed it's in the form of a loan you have to pay back with interest with an ISA and its repayments as collateral, or it's a sale to a neutral SPV with recourse in the case repayments don't hit a certain threshold. In the rare instance (we've never done that) schools have been able to sell ISAs full stop, it's been at extreme discounts or based on discounted predicted likelihoods of future revenue, and if those ISAs don't repay the buyers bail and the school trying to sell them is out of business.