The Automated Workforce - how AI will disrupt the developer’s job. This was the title of a conference talk I gave in 2018 about whether advances in AI are about to make software developers’ jobs obsolete. This was before ChatGPT had entered the zeitgeist; before Sam Altman’s return at OpenAI (along with some “business-friendly” board members) effectively destroyed any claims to altruism that company made; before Timnit Gebru tried to bring ethics to Google’s AI research and was punished for it. LLMs weren’t in the public consciousness yet, but there was some serious ink already being spilt in the media about what this coming technology wave meant for all of us.
Now I am in no way an expert on GenAI; but I was strongly motivated to do this talk for two very important reasons. First, it was a chance to speak at a conference alongside Martin Fowler, and second, there was a free trip to New York City involved.
The germ of the idea was simple; so common-sense that it made me nervous to try and sound intelligent on stage about it. It said “we are nowhere close to the AI singluarity; this technology is just going to do is bring efficiency to the most mundane parts of our job. And efficiency isn’t something to be feared if you’re tackling large complex domain problems. Embrace it, adopt it, and keep growing.” It wasn’t too convincing, to be honest. Neal Ford, in the audience at a practice session the day before, told me “You have a third act problem.” And right he was.
Historically, the software industry has been the beneficiary of technological progress, never the victim. In order to benefit from this coming AI revolution, we need to embrace it early: through investments in data science literacy, DevOps, and platforms. At the same time, we will need to tackle some long-standing problems in the tech industry: ageism, sexism, lack of diversity & inclusion. Because when development is defined not as “coding” but as “problem-solving”, teams that bring a diversity of ideas, experiences and viewpoints will have a crucial competitive advantage.
To my great surprise, the talk was quite well received. Mainly because it had more pop culture references than all the earlier talks that day, which were far more technical but also bone-dry. Audiences like to be entertained, not lectured at. So I wasn’t a complete dud. Still; I always regretted that I never did justice to an interesting subject.
Enter the master. It takes an expert communicator to take a complex idea and convey it in a clear, entertaining way. Enter the legend that is Bryan Cantrill. In his recent talk *“Intelligence is Not Enough”, Bryan takes on the idea of what true “engineering” actually looks like; and why AI is nowhere close to replacing us in that domain. And in his inimitable style, he reminds us of what “human” qualities we sometimes fail to value as engineers. This is the talk I wish I had given; the full fruition of the germ of that idea. Take a look, I guarantee you won’t be disappointed: