tl;dr this is just another rant on the state of technology, a babbling about enshittification.
So, here I was, walking next to the bedroom door: “oh crap! The light is still on.”
I approached my Nanoleaf and pressed the power button… and the Nanoleaf still shone, as clear as the day.
I sighed, but “hey you got a computer in lieu of a light, have you tried turning it off and on again?”.
So I knelt to the gods’ will, and also next to the electrical plug. Unplugged, replugged, and it took me about a minute to assess the state of it. Because… a computer also has a boot time.
Well… my stubborn and overpriced smart led strip, that never failed to disappoint, was dead.
The quite reactive Nanoleaf service, after a few extra manipulations, sent me a fresh component, that still refuses to connect to my phone or wi-fi, but at this point I was just asking for my bed light to be back. And it was. The hours of my life dedicated to trying to get this to work weren’t, they never will.
Hardware
Hardware problems keep me up at night. Especially this load of internet of shit that the industry keeps spitting out because an obscure board can not get over being passed the hype and needs badly to follow the innovation imperative.
- Hello “smart” LG washing machine. I never managed to open an account on your service to launch the app.
- Good day “smart” Siemens stoves. You had to be wi-fi compatible for every non entry level models… I still do not know why.
- I am not forgetting you, not so smart Siemens microwave that has a two seconds lag (if not already on) before opening its door. Because again … a system has to boot when nobody pressed a button in quite some time.
- Greetings to our Philips Hue “smart” lights at the office. You work quite ok, except when the lights stay dimmed for no reason, or in this bathroom stall where the motion sensor refuses to work.
- Namaste Nuki “smart” lock system. You are awesome, except for the building downstairs door that seems to only work when our phones have enough network coverage.
- Nobody claims that you are “smart”, wireless Sony headphones, but hey … we all know that Bluetooth is a half-assed job made a world standard.
- You are just the headphones little brother, dear Keychron K2 keyboard. It is always a pleasure when you interrupt a productive work session just because of a low battery that does not give any warning or because the remapping of some keys is suddenly off. I have no proof, but my instinct gives some credit to macOS on that last one.
- I have not forgotten you lovely Google Nest Cam, I know you are here, still plugged where I left you when I grew tired of the constant changes about how I am supposed to use you.
Ironical thoughts set aside, the only thing that seems to work quite ok is my Roborock vacuum cleaner, whose only drawbacks were to absolutely need me to have an account and have once “eaten” and stripped an electric cable. At least, this thing saves me time.
SaaS
Unlike what you believe the initial meaning of SaaS is not “Shit as a Service” but “Software as a Service”. I also like “Software as a Sévice” ; “sévices” (always plural) meaning “abuse” in French.
About a decade ago, I recorded a series of podcasts (called radio shows at the time). The aim was to bring bits of technical understanding to an entrepreneurial audience, one of them was about SaaS.
At the time, open source solutions were flourishing for a lot of end-users products: project management & source code hosting (Redmine, Trac), communication (BigBlueButton), CMS (Joomla, Wordpress), CRM (VTiger) and many others. Installing those software was usually a major source of pain, some of them were half-baked product and the maintenance was painful. SaaS seemed to solve, for a fee, a lot of those problems. For the people providing such solution, it was also a way to make their business more money generating and because it was viable the solutions were often better on the long run.
But as I am a dedicated person, I remember sitting at my kitchen table trying to also find the negative aspects of this revenue model. Consumer captivity was definitely one of them, in comparison to open source: you do not own the data, you do not control the software. Migration is, by design, a major paint point.
Once you, the customer, are locked in. Maybe a feature will disappear, surely the prices will increase. Sadly too, you might live on a niche segment and the features might legitimately evolve targeting a larger segment. Or you are just a figure in a quarterly reporting, and it saddly happens that making redesign #132 has been seen by the project manager as the best way to advance his career (I am looking at you Google). When seeking the next euro of EBITDA, current customer satisfaction is usually only measured by churn (departure of customers). So as long as the barriers are high enough every move will feel like a demonstration that the users and makers interest are not aligned.
Making software
You can attract people to buy hardware and SaaS by over promising and not respect yourself enough to deliver. But the lack of self-respect has an aim, it is financially better.
However, strangely, the same problem appears when making software but the financial incentive is not as obvious. I think it is hidden behind career building and not questioning the hype. The term of Resume Driven Development seems to have been coined to describe this.
There are a few symptoms that would name here:
- The disregard for backward compatibility.
- The appeal of “new” over “old”.
For me, at its core, the promise of software is “solve a problem once and for all”. Software does not need oil change, it does not tear with usage or time. Security patches are a topic, it would be a lesser one if backward compatibility was not a remote idea. It seems that there are still strong drives towards obsolescence that achieve nothing but burning man-days at the stake.
It translates in:
- Disappearing API or interfaces.
- Breaking changes.
- Package managements systems that sometimes feel like they only have been designed for the last version of each managed packages.
- Inappropriate use of technologies because it sounds better to say “Mongodb” rather than “MariaDB” or “PostgreSQL”, or “GraphQL” rather than “REST”. Old, working, boring technologies? Urk. The cure to this disease is clearer, it even has a website boringtechnology.club/.
- Technology lacking maturity being incorporated into products because of hype pressure. If your industry has had chatbots in the past I am sure it has already migrated to LLMs, hopefully with better results.
One peculiar example I have of this is my usage of AWS. I bought the cloud services’ promise that I may not have to care about setting up servers anymore. But often, when I am deploying a new system, my old routines and utilities do not work because “this API has been retired you should use version X.X.X”. The promise is fulfilled, I don’t care about updating my servers anymore, I care about updating to their ever-changing APIs.
Does this even work?
I am describing a vision where everything is feeble but “hey the world is all running for the better, right?”
Is it? Most technologies are not really end user products. The only end user product is an experience that puts the product possibly along with others in order to build a consistent and reliable experience, hopefully meaningful in at least one person’s life.
Now tell me:
- If I am gifting you a smart™ product, how joyful do you feel at the idea of setting it up? Would your mom feel the same?
- If this product asks you to install an app, what is your immediate feeling?
- If this app asks you to create an account, what is your immediate feeling? Besides the immediate “I knew it” moment.
- Among the apps you use weekly, how many crash daily?
- If any, how many of those likely generates a few millions of revenue a year?
- How often does it feel that you have to fight against the products you own and paid because they want to coerce you?
- How often does it feel that you have to fight against apps because they want to coerce you?
Conclusion
While this was fun to write. It is mostly a rant, I don’t have a silver bullet nor many constructive remarks.
It comes down to the chinese concept of Chabuduo (Hu Shih, 1920) which means “close enough” or “good enough”. “Good enough” for market traction, “good enough” to generate sales, but “good enough” is not good. Most experienced consumers are weary of trying new things, they prefer something old that work to something new that only may work better.
It reinforces the idea that in order to gain a market you need to reliably promise a 10 x improvement (Zero to One by Peter Thiel & Blake Masters). However, we would all benefit from marginal form of enhancements, if real, not overpriced, nor overblown: a Siri assistant that actually works, a Remarkable tablet that does not lag, a short range wireless communication standard that does not suck.
