On Being Indispensable

Busted up at this point.

But seriously, this is me, though to a lesser degree (this guy sounds much awesomer then me).

10 years ago, I exited a shrinking technology community. I loved it, but the signs were on the wall. I knew I’d be one of the last ones to turn out the lights. I was just past 41. The prospect of looking for work at 50+ with guru expertise in a Cobol like technology terrified me.

There seemed to be two basic paths at the time. Embedded or web. It seemed like web development was getting commoditized, community college grads earning $17/hr to hammer out a web site for the local pet store. Embedded seemed more lucrative, and I had a decent level of C competency.

I made a choice to become a polyglot as well. I would embrace the “best tool for the job” mantra, and avoid being pigeonholed by any single technology.

An opportunity in our small community opened up to be somewhat entrepreneurial for a medium size company and I jumped in. Our number of participants is no where near as large as the article. 10 at most. But 10 years later, I’m indispensable. I own the embedded C code that runs the 3 different node types on a proprietary LoRa network. The protocol that communicates with the edge device that runs embedded Linux, running multiple systemd services, the uboot configuration, the cooperating Python programs, our own BLE driver from said Linux board, the MQTT and BLE communications schemes and binary json like protocol that communicates to apps. The Kotlin code that runs the two apps. The horror that is working with BLE on Android. The objective-C then Swift code that runs the same two iOS apps. The elixir service that transforms MQTT traffic to swaggerified API. The ansible scripts that configure our servers. Etc. I try to document things, write good code, unit tests, but there’s still a huge amount of “how it all fits together” and “we tired that, don’t want to go there” knowledge in my head.

I have enjoyed learning all these things and more. But I regret I never get the time to get really good at any of them. I regularly experience tuple dysphoria (“how do we do tuples in this language again?”) and other similar “why do all of these languages need to do the same thing differently?” And feel like I always probably need to go figure out how docker works or some other new thing so that I can keep pushing our product offering into new and exciting places.

And like the fine article, I am well remunerated, I am indispensable, I don’t know if I can keep this up til retirement, and it’s kinda lonely. I don’t see a way out. People want to pay $120K for a dedicated Elixir developer, not 180 for “does a lot of fricking things kinda.”

A couple years ago, we hired someone to “help Travis” but it didn’t work out so well. I’m still trying to figure out why.

(I should add that there’s still interest in hiring another body to participate in this madness of polyglot indispensable-ness, if such a thing sounded appealing, email’s in the profile)

No such thing as "indispensable". Very few companies and projects die to a singular departure, you're just not that special kiddo.

What's actually happened here is a selfish desire to know things without sharing them. If you learn something new, share it somewhere.

"People came to ask the Oracle questions only the Oracle could answer for the Oracle refused to share knowledge by any other means".

Confluence is pretty good if you use it properly. Weekly knowledge share sessions are also handy.

Learning a thing and sitting on it makes you an asshat, not an asset. Seems OP learned the hardway, but they still need to remove their ego from the equation.

You're not good at knowing things (basic human capability that), you're terrible at sharing them.

The méthode en vogue of the current era is to jump ship every 2 years because that's the best way to get a raise/promotion. With that strategy being indispensable isn't that important because you will self-dispense in 2 years anyway.