I'm not holding my breath
Technically, there's simply no point. Cross-platform portability just isn't that hard for code that only processes data. Any app that performs device control will need some kind of native glue to the device, but not much (since many devices are just "our protocol over a USB link" or something similar). Most apps will need some sort of file and directory management, but that's just a few hundred lines of OS wrappers for each platform and has been done a million times before. The sockets API is supported on any platform that actually wants a non-zero market share, and through that API (and numerous RFCs and similar standards) you have myriad additional services in a standard form. The rest is your own algorithms.
The only domain where this argument falls down is the UI and whilst I might forgive the general public conflating "end-user-visible software" with "all software", it is rather disturbing to hear a CTO make the same mistake.
The GUI is a sticking point for portable software largely because we just didn't agree on how to do it. We had several different approaches that worked (X11, Mac toolbox, MsWin) and various attempts to make each available on other platforms. As APIs, they had strengths and weaknesses, but there was no clear winner. To be honest, we still don't have a clear winner in terms of capability, because HTML5 is a strait-jacket to anyone familiar with a native GUI API, but it offers "adequate for many purposes" in combination with "fully portable" and perhaps that will do.
And in any case, aren't we all going to be re-inventing the whole field of software engineering once we figure out how to make GPUs do anything except embarrasingly parallel number crunching?