I’ve been around long enough that I normally roll my eyes when ‘the new shiny thing’ shows up. If I start gushing about something, I truly see it as game-changing in a market space that is full of things that simply are not. Perhaps an example is in order. For years after college, my hobby was compilers and linkers. Until my Master’s program had a class in compilers and the newer variants of Lex and YACC showed me that those skills are archaic. Lots of fun, and I’ve kept my source to a 100% original MS-DOS linker, thinking one day I’ll update it to generate modern executables. But I won’t because the other thing that reduced the appeal of compiler and linker skills is open source. I can download code that, by virtue of being exercised by like a billion people, is inarguably better than my efforts were. Better to put my attention elsewhere. Open source was indeed a game changer. Today, when I see there is a new language du jour, my very first question is, “Okay, but why?” the answer is rarely, “To make developers’ lives easier.”
The same is true with generative AI. Just as open source has spawned a never-ending collection of failed or misdirected OSS projects, AI is starting to create buzz about dozens of “that’s not really useful” use cases. But amongst all those are the “This changes everything” use cases, and for IT–be they Devops or devOps, generative AI for code creation is huge. This will almost immediately become a mandatory job skill, and you should start learning how to get generative AI to spit out the code you need today. It will increase your productivity, help your company, and it will increase your career prospects, helping you.
Generative AI Goes Further
Two examples. In completely different projects that aren’t even for companies in the same space, Lori and I both used generative AI to develop systems faster. I know very little about Lori’s work other than that it was to generate Python that did the grunt work and took care of the repetitive bits she hates. My work was to generate bash scripts for things I don’t generally do in Linux. Any honest assessment agrees that the power of Linux is masked by the insane level of obscure commands. While researching online has helped us get over the “What the heck does that do?” for decades, generative AI goes one step further. It allows you to tell it what you want to do, then give it whatever additional information it might need to tweak the results, and the script is done.
We all use APIs, libraries, SOs and snippets pretty much daily; this is literally just an extension of that usage. Most of us would search online for examples of what we needed to do, then copy and modify them. This really is literally, “Give me an example of that. Good, now customize it to my needs.”
There are risks with any significant change, and generative AI for script development is no different, but the largest risk identified – that entry-level people will not learn and grow – is blown out of proportion. Given an example to read, the curious and those who are going to be responsible for supporting the system will go over it with a fine-toothed comb. The learning will be more reading than doing at that point, but in the end, the script still has to execute. Troubleshooting won’t perform itself, and people will continue to strive to understand — both because they need to when supporting systems, and because curiosity is indeed the hallmark of IT staff.
You all are the backbone that keeps the organization online and employees productive. We are way past the time when 90% of companies would fail without IT. Keep rocking it, but give your org that competitive edge. Take any tool that will increase your productivity and help your career in the process.
