Since it’s launch, the automatic updates “feature” has been a real pain for Windows 10 users. It has the tendency to force itself on users and launch updates at the worst possible time.
Shure, Microsoft has introduced snooze features last year to try to alleviate the impact of forcing updates on users, but that didn’t completely prevent that suddenly, you found yourself in the middle of a 3-hour update, just as you were going to head into that career-defining meeting.
Today, Microsoft announces that they have come up with a new way of avoiding frustrations while at the same time ensuring that your machine stays up-to-date: Machine learning.
How does it work?
In Microsoft’s own words:
Have you ever had to stop what you were doing, or wait for your computer to boot up because the device updated at the wrong time? We heard you, and to alleviate this pain, if you have an update pending we’ve updated our reboot logic to use a new system that is more adaptive and proactive. We trained a predictive model that can accurately predict when the right time to restart the device is. Meaning, that we will not only check if you are currently using your device before we restart, but we will also try to predict if you had just left the device to grab a cup of coffee and return shortly after.
But is it failproof?
We’ve been using this model on internal devices, and we’ve seen promising results upon rollout. Due to the nature of its architecture, we’re able to update the model with minimal turnaround time based on our insights from its performance. It’s all thanks to our cloud infrastructure.
Sound cool, but as with all things Microsoft and Windows, whenever you see the words “try to” it means that they expect the algorithm to “mostly” work, but won’t get it right every time. And if that time is when you most need your computer, the whole experience will feel as horrible to you as it felt before.
Is there a good approach?
Yes, there is, in my opinion. Let’s draw an analogy to the car world. If your car needs urgent maintenance, the car’s computer will inform you. It will insist with bleeps and red lights if things get critical, but at no point, it will just shut down in the middle of the highway because you refuse to listen. Unless of course, there is a mechanical problem caused by the lack of attention you gave to the clear warnings.
It also won’t autonomously drive to the garage to execute the updates/changes on its own. And this is a critical environment, where mechanical problems can cause fatal accidents if left unattended.
What it might do, however, and especially if you own a recent model, is to refuse to start to prevent further damage. This is sensible, as it forces the owner to act, and it can prevent a disaster.
Back to computers. Unless we are talking mission-critical situations, such as operating rooms and transportation (and I highly doubt Windows 10 is used for critical tasks in such environments), a Windows 10 machine that boots unexpectedly won’t kill anyone. So why not force users to update by increasingly preventing them for working? For example, if you don’t allow a downloaded update to run after 5 prompts, severely limit the internet connection speeds to minimal service. Or lock the screen’s brightness to 5 pct. Or limiting the CPU’s speed.
Stuff like that. That way, users will end up allowing the update to run to regain full functionality, while still allowing them to properly finish up what they are doing.
This could be complementary to the machine learning, that can jump into action when the system is 99 pct sure that it’s the right time to update. In the cases where there is more doubt, the system can decide to annoy the user by limiting some core functionality, with the clear message that it will be restored once the update is done.
But why not let the user free to decide? In 2018, the risk is too high. An OS needs constant updates to stay ahead, both functionality and security wise. So going back how things were done in 1999 is not an option. If Microsoft can find ways to more acceptable for their users, they can only benefit from the increased user satisfaction they will get.
Anyway, that’s my take on it. Comment below on yours.