Is the age of Evergreen Apps coming to an end?
One interesting thing about software developers in general is that most of us aren’t even aware of the fact that all we’re doing is making…
One interesting thing about software developers in general is that most of us aren’t even aware of the fact that all we’re doing is making so-called ‘evergreen apps’.
So, what’s an evergreen app? In a nutshell, it’s an app that was built with longevity in mind. To be able to make it and keep it evergreen (meaning, last forever and always remain fresh), we strive to make it as generic as possible.
For example, let’s look at the familiar old staples such as a word processing app, or a spreadsheet app. Those apps were designed to be as generic as possible. As such, they have survived many decades of real life usage. What’s their secret to longevity?
Basically, a generic app is an app that offers exact same functionality and exact same experience to any user. If, as a designer, you don’t have to tailor the app to any specific user’s needs, it is easy to make it in such a way that it will remain kind of sort of useful for many years.
Concierge apps
I call those generic apps ‘concierge apps’. Why concierge? Basically, the experience one gets when using such apps is similar to the experience one gets when interacting with a concierge. Simply put, a concierge will give canned, prefabricated answers to anyone who approaches them. You know, you approach a concierge at any event (such as a conference) and they will explain to you what’s on the agenda, where are the coat checks, when will the event start, where are the facilities, etc. Anyone else who approaches them will receive the exact same answers.
Similarly, concierge apps give identical menu options to all users. Also, regardless of how many times have you already used the app, your experience remains the same. It is generic, faceless. Bottom line — same as with real life concierges (who represent the organization), concierge apps represent the company that has built and marketed the app. Neither real life concierges nor concierge apps care about the individual specifics, quirks, and foibles of each of its clients.
Butler apps
In contrast, we can envision a new brand of apps that I like to call ‘butler apps’. Same as with real life butlers, butler apps represent the needs of their employers. Butler apps do not represent the organization, they represent the individual human user.
Such apps are designed and tailored to learn about the specific individual needs of each of their users. Naturally, it is not difficult to envision that designing such apps would be many orders of magnitude more difficult than designing faceless, generic apps.
To put it differently, while generic concierge apps resemble mechanical devices, butler apps cannot be designed without imbibing them with certain level of intelligence.
Yes, but how to enable butler apps to behave intelligently? Enter Generative Pre-trained Transformer (GPT). Right now, GPT seems to be taking the world by the storm. And while there is plenty trepidation about this emerging technology, the fact is that we have opened the Pandora’s box, we have opened the can of worms, and we have let the cat out of the bag. There doesn’t seem to be the way back into the age of innocence. So, we might as well look into how to make the best out of the newly emerged situation (or, as the Russian proverb says “don’t shut the stable door after the horse has bolted”).
In the human world, daily affairs are intricate and complicated. Aiming to design an app for each of the myriad unique scenarios is a recipe for futility. But if we are to graduate from making evergreen (i.e., concierge) apps and start making butler apps, it seems that we have no choice but to begin crafting very narrowly specialized apps. Those narrow focus apps would then be configurable to learn from each individual user and adapt to their quirks.
Mission impossible?
It doesn’t take a very deep insight to realize that the above plan is unattainable. The cost of hand-cranking software apps is already prohibitively expensive, especially if we were to start creating thousands (if not millions) of narrowly specialized apps. So, does that mean that hypothetical butler apps will remain a nice idea that will never see the light of day?
Have we embarked on a mission impossible? I’d agree that we have, if it wasn’t for the GPT breakthroughs. I see GPT possibly coming to the rescue. In what way? GPT is rapidly becoming a problem solving technology. We give it a challenge, and watch it churn its way toward some kind of a solution.
That behaviour is similar to the behaviour we observe in good butlers. Their employers are posing daily challenges, and it is butler’s job to figure out the solutions.
And as a good butler focuses their entire attention to the behaviour of their employer, such butler becomes adept at being the first to understand the situation, the first to adapt to it and respond, and therefore the first to overcome the challenges.
In behaving that way, a good butler turns mission impossible into mission possible.
Single use apps
Evergreen apps are always envisioned and designed to be used over and over. Hence the focus on longevity (it costs a lot of money and time to craft software apps, so we’d better get our money’s worth by using them for many years to come). And that arrangement works fine for generic scenarios, that is to say, for situations where it is not important who is using the app.
But in the narrowly focused scenarios, where each user is most likely going to expect different functionality, such one-size-fits-all apps are not only useless, they are positively annoying. It is for that reason that we now seem to be moving into the area of so-called ‘single use apps’.
What does the term ‘single use app’ denote? It basically means that, as some situation emerges, and the user needs a solution, the desired solution is crafted on the spot. The user then leverages the newly created automated solution, and once the user’s goal is reached, that newly created automated solution is summarily disposed of.
Why such wasteful approach to automated solutions? The reason is that the solution is very narrowly tailored, and won’t be of any use to anyone in the future.
GPT holds promise of being able to craft such single use apps. Being a good, dedicated, loyal butler, GPT can learn things from its employer, and will then be able to devise, on the spot, automated solutions that will please its employer. Those solutions will be produced on the spot and will therefore come at a tiny fraction of the price it takes to devise an evergreen app.