Say you decide to code up a bunch of risk analytics in Scala. Say Scala doesn't do it for you 2 years from now. You really haven't lost anything - just hire a bunch of VBA jockeys & they'll migrate you back to Excelland in a reasonable span. This isn't complicated device driver code where module A invokes module B with assembly address C...More like convexity on the 10 year swap is different from 2 5 year swaps & that difference can be exploited with a bigsize bet on payoff X and a ratio hedge in case bad things happen. Whether you do all of the linalg/stat/regression scenario analysis in Excel or call a Mathematica routine from Scala should be irrelevant, you get the same answers. So I simply think the risk-reward ratio is out of whack. You're willing to take huge risks on stochastic variables 10 years out into ther future but unwilling to touch a programming language/platform because its too new, even when there's a clear-cut migration path. Seems silly. Don't see how that attitude translates to being cynical about finance.