Just say no (to proprietary SaaS)
2025-11-13
Some things advance too quickly for us humans to intuitively keep up.
For example, our brains interpret sugary foods as tasting good despite them not exactly being full of nutrients. The level of junk-food firepower we have access to today is completely unnatural. Our brains haven't had the time to "realize" that these foods are actually harmful and therefore they should probably taste bad for our own good.
Or consider drugs that overwhelm our dopamine reward system in ways that aren't possible in nature otherwise. Our bodies are tricked into craving those experiences again and again which is called addiction. Since drugs are a brand new invention in evolutionary time scales, we can't rely on our primitive instincts to avoid their harm.
Luckily as humans, we have some higher-order thinking that we can apply to these situations to make up for when our instincts fall short and tell us to just do the destructive thing that feels good.
Now that the analogies are sufficiently primed, let's talk about cloud computing.
Proprietary SaaS is a deal that's too good to be true
By now, many people are aware of the shift occurring as software transitions en masse to the Software as a Service paradigm. Or should I say transition back as this model is conceptually similar to the mainframe/terminals of the 60s and 70s.
SaaS today offers considerable benefits over the more decentralized computing paradigm made possible by groundbreaking hardware advances through the 80s to 00s. For example:
- SaaS is the easiest way for users to consume applications (no installation/configuration required)
- publishers can manage the software lifecycle (updates) more easily which usually leads to a better user experience
But, as you read the title of this section, SaaS in practice is usually partly good and partly ominous.
SaaS offers enticing features that we couldn't get before at basically no cost. How could you, before the cloud, effortlessly collaborate on documents in real-time and be immune from data loss? It was simply impossible when we were sharing files via CDs back in the day.
If this next generation of computing is better than the last, shouldn't it at least cost something? The last time I bought an improved piece of hardware, I had to pay money for it. If we suddenly have more AND it's free, that should trigger some concern. If you asked a random person from the 90s, it's a deal that's too good to be true.
Intuitively, this is a contradiction. But, since it's free, we take the deal anyway. And, since no harm was overly caused, we take the deal again next time. This is how SaaS became normalized.
The way this makes sense economically is, of course:
The users are the product
It's such mainstream knowledge now, it feels cliché to actually write it down. The reason SaaS can be offered for free is because of the enormous value of our collective data and the enormous value of our attention. The features are really just bait to harness users, either unaware or apathetic of the actual deal they're taking by using certain products.
Companies are heavily incentivized to design their products as SaaS now. It literally makes their valuations multiplicatively higher. Which means companies are going to keep pumping out SaaS as fast as they can.
Since there's so much complexity and abstraction in today's software, we can't easily tell when we're being used either. This stuff is too new for our intuition to have developed adequately.
It's all about control
Control is power and power is control.
Software companies/developers already had some level of control simply by virtue of having created the software, but SaaS elevates it to a level hardly possible before. Previously, you could expect software to generally protect your interests and privacy. Today, companies can't resist exploiting users' privacy because it's become too easy and too profitable.
If you don't control your software, "your" software will control you. You now have to be intentional about protecting your computing experience.
Cars without steering wheels
It's inevitable that self-driving cars in the future won't even have steering wheels. We will lose the ability to control exactly how our cars go, despite telling them exactly where to go.
The big upside is that computers are much safer drivers than humans. Especially when all cars on the road are driven by computers communicating in a peer-to-peer manner. We'll be both travelling faster and dying less in car accidents by giving up manual driving, but we'll have to give up some control over our transportation to get it.
How much control do we want to maintain? What if we can't drive anywhere without it being logged? What if the car is hacked and drives you into a tree at top speed?
When there's a clear win like "dying less in car accidents", I think giving up control is the right thing to do. Transportation is really just a means to an end after all.
I can't find the analogous "win" when it comes to SaaS. Sure, applications can be delivered and managed more easily than before which leads to a better user experience, but compared to the "users are the product" downside from a couple sections ago, it's not much of a "win".
Computerized cars
We don't have to imagine a future where manual driving is outlawed as too dangerous to see the loss of control of our cars.
When I bought my first nice car, the salesman hit me with the obligatory warranty pitch. It went something like this:
Salesman: "I wouldn't buy a car this expensive without a warrany. What if the wheels fall off?"
Me: "I like fixing my own cars." (having previously owned an old Jeep, I was experienced in fixing all kinds of things myself).
Salesman: "But you can't work on these cars yourself - they're all computerized."
Me: "Well it's good thing I'm a CS major then." (realizing I was screwed if anything went wrong with the car).
I succeeded in dodging the warranty, but in that moment I realized I had less control over the new fancy computerized car I was buying than my old one.
Despite fully paying for the car, there was still some sliver of collective ownership/control vested in any auto shop capable of fixing it.
On the bright side...
So far this has been thoroughly doom 'n gloom, but there's light at the end of the tunnel.
Usually when an impactful technology is invented, there's an adjustment period where things aren't quite right. It can take time to work out all the bugs - not just the technical ones.
For example, an untold number of people had to die after cars were introduced before we figured out seatbelts, airbags, and crumple zones.
By caring about these issues, we can collectively set a new standard for ethical SaaS that respects user freedom and privacy.