I actually take very few exceptions to what Tim has to say in that video. I believe he is bang on w.r.t. multi-core, REST, (lack of) RIA, the strength of HTTP and the lack of evidence supporting a need for push-based applications (i.e. the argument that poll-based doesn't scale is completely invalidated by the fact that it does as RSS, twitter, email and a host of other very large scale apps are poll-based).
In the past 2 years, I have spent 3490 hours on the "business at hand" [...], 10 hours on infrastructure (setup and maintenance of trac, svn, backups, VPN)
This is very much not inline with my experiences. In most organizations I've been with, the "build master" is a FTE. The "server guy" is a FTE who maintains dev and QA machines (though may also maintain a few other corporate servers). The docs team spends a whack of time maintaining the documentation system. The bug tracking system is a load of garbage because no one is willing to spend time on maintaining it...so emails and/or spreadsheets and/or poorly implemented OSS tracking systems are used thus wasting time and resources because it cannot support a proper SDLC.
All of this is wasted resources. For the majority of companies, these should be commodity systems and thus could be outsourced to someone else to maintain.
Choice of outsourcing should be based on reducing resource cost on the development team, not on reducing the price. Development resources are worth WAY MORE than their hourly wages.
I would never ever switch to a remote {webhosting|email|CRM|ERP|backup|datacenter|...}
Many, many times in history we've seen this initial gut reaction to the idea of "remote data". And many, many organizations are recognizing the business benefit of no longer hosting/maintaining their own infrastructure.
Focus on the business at hand (e.g. coding) and quit wasting time on infrastructure (version control, defect tracking, build systems, backup & recovery, server sizing, etc...).
I don't currently foresee our organization moving to remote IDE, but if we decided to cut costs on non-core areas, outsourcing IT infrastructure would certainly be one option (we've already outsourced a portion of it).
Luckily for us, science always corrects itself in the long run.
Are you talking about the Darwin effect ?
Turning the world a billionth of a degree, probability drive, the end of the universe, Marvin's vast level of intelligence (though self-claimed), music derived from stock market trends...all ignited (or rekindled, or simply increased) my love of mathematical constructs.
Real Programmers don't eat quiche. They eat Twinkies and Szechwan food.