What I have found is that this is often made up for by the arbitrage opportunities! Because traditional industries are so out of touch with the latest hipster

What I have found is that this is often made up for by the arbitrage opportunities! Because traditional industries are so out of touch with the latest hipster trends, giving them just a little taste (calibrated to appear as bottom line-relevant as possible to the right stakeholders) can make you seem like the second coming of Jesus, get you massive raises and promotions, and make you essentially layoff proof. I remember I had a job in 2006 - where I was a low-grade operations nobody - and I prototyped something in client-side JS for an internal project. Overnight I became Master Sensei, and everyone in engineering (not my division) knew my name. They were mortified when I tendered my resignation. It doesn't always work. Megacorps don't intrinsically care about new technologies for their own sake, and it is true that defending the value of engineering to a bunch of greying MBAs in financial services or retail can be demoralising, at times infuriating or heartbreaking. Sometimes this type of activity can be consigned to skunkworks that never get a hearing or see the light of day, whether for political reasons (you're infringing on the J2EE guys' turf) or because management isn't tech-savvy enough to understand the value. But man, when it works, it works! I'm in telecom, and much of the industry - by the numbers, forget Twilio and WebRTC - operates according to soul-deadening customs that have changed little since the 80s and the advent of the micro-computer. Core routing tables and key provisioning databases are still distributed as giant spreadsheets and Access databases. Cable companies can be surprisingly low-tech, too. I've seen what happens when one becomes their connection to spiffy, modern innovation techniques like regular expressions and Python ETL jobs. The threshold for mind-blowing Bushido is startlingly low in enterprise. That can really make you an A-player in their eyes. Your main competition is going to be that these companies prefer to source such insights from giant professional services companies and consultancies, who can be similarly low-tech, but some of whom have realised there's enormous money to be made in opening large companies' eyes to 2003-era technology like Perl and Subversion. There's also the fact that big companies don't like the risk profile of depending on one guy or a small team to do advanced stuff. They'd rather spend millions of dollars to have it done mediocrely by Accenture or Deloitte, in the customarily obese enterprise fashion, skies darkened with "Linux resources" and the Sun eclipsed by UML diagrams and process maps. The other main danger is, as you allude, falling behind because the threshold for being ahead of the curve is so low. If you show them screen scraping or write an internal portal in a modern JS framework, they'll think you're from the future. Then you get into an actual tech company job and realise you've become an entrenched maintenance specialist in some decade-old stuff.