This article was originally published in my weekly IoT newsletter on Friday August 18, 2023. It was my last newsletter.
I wasn’t going to write anything more in this newsletter, but I was feeling inspired and so decided to leave y’all with one final opinion — and one final mandate.
During a webinar Thursday focused on how journalists can use generative AI, one of the panelists bemoaned the fact that developers in the newsroom don’t want to invest in building AI verification tools because the problem of detecting deepfakes is constantly evolving, meaning the tools would need to constantly evolve. I instantly empathized because for years I’ve been bemoaning the constant maintenance I have to put into my smart home, my phone, and my other tech services to keep them up to date.
This goes beyond applying security updates and involves swapping out home automation rules when one provider changes its APIs, or figuring out a new workflow when a link between Zapier and one of my digital services gets shut off. I’ll even include the moments of confusion that arise when I get into my car and find that Tesla has issued a new software update and changed the position of information on my dashboard. Living with technology means living in a state of flux.
Instinctively we know this, but operationally and mentally we have yet to adapt to the need for constant change. Mentally recognizing that a connected device is valuable because the feature set can continuously expand is one thing, but getting frustrated when the UI changes or a new recipe setting on your smart oven cooks your chicken differently is another. Most of us aren’t equipped to deal with constant change in all aspects of our lives, so the friction when your car or oven works differently is jarring.
Operationally, businesses have yet to adapt as well. Google provides a very good example of this. Google rewards its employees for creating new tools and software, which means that if you want to get ahead you need to invent. This leads to a culture where employees are incentivized to innovate as opposed to maintain. But as users, this means your favorite tools are in a constant state of decay or that when partners make changes, Google may not react as quickly.
Google isn’t alone in valuing innovation over maintenance. You can see the clash in industrial settings when the IT staff accuse the operations engineers inside a plant of creating a culture of no, refusing to add automations or new technology to the established way of doing things. But one reason those OT staff are hesitant is because ops engineers have a decades-long history of building up processes that are consistent and knowable. Adding IT to the mix adds entropy and a need for new maintenance procedures that IT can sometimes balk at.
But while the OT side has an established culture of valuing maintenance, the IT side does not. And that’s what we need to move toward as we embed computing in more and more devices and processes. Software decays. And it does so rapidly.
Software’s inevitable decay will continue to have more and more impact on our day-to-day lives, so we need to incentivize developers to maintain as well as innovate. We need to give employees time in their work weeks to adjust to new user interfaces or changes in their services. We also need to leave them time to play around with new tools and advances, and recognize that as productive work.
Much like law firms and doctors are required to get continuing education, any profession that regularly deals with technology (which will soon be most professions) needs a similar ethos that both incentivizes and supports employees as they try to adapt to continuous change. This includes, but is not limited to, the advances posed by artificial intelligence.
And yes, it will take a more curious and engaged workforce, but it can’t be all on employees to muster up that energy and enthusiasm. Maintaining services and adapting to changing software and services will become an economic value and will need to be treated as such.
On the home front, it may mean that consumers subscribe to a product over time to ensure that the company behind said product pays the ongoing developer costs needed to maintain it. Or perhaps it means a company sells a product with a set expiration date for when it plans to stop maintaining it.
If we want to add intelligence to our everyday products, we can’t simply focus on new features and innovation. We have to think about how to maintain those products, and how to pay for that maintenance. And as computing enables more software and services to infiltrate our workflows and more jobs require the ability to embrace new innovations, businesses need to invest in maintaining their employees’ skills.
Constant innovation is exhausting. And because that innovation is currently based on software, it’s prone to rapid entropy. We have to value the people and the time it takes to counteract that entropy while also ensuring employees and consumers have the space to adjust to this change.
And now, I’m off to go make some changes of my own.