(These are excerpts from my book "Intelligence is not Artificial")
A Consumer's Rant Against the Stupidity of Machines: Reverting Evolution?
When you buy an appliance and it turns out that you have to do something weird in order to make it work, it is natural to dismiss it as "a piece of garbage". However, when it is something about computers and networks, you are instead supposed to stand in awe and respectfully listen to (or read) a lengthy explanation of
what you are supposed to do in order to please the machine, which is usually
something utterly convoluted bordering on the ridiculous.
This double standard creates the illusion that machines are becoming incredibly smart when in fact mostly we are simply witnessing poor quality assurance (due to the frantic product lifecycles of our times) and often incredibly dumb design.
For mysterious reasons, the computer industry has a tradition of being hostile to its customers. Buying a new refrigerator means that you will have an appliance that saves energy and provides one or two useful new features. Buying a new computer, instead, typically means that some of the important things that you used to do every day will now become very complicated or just impossible. A whole generation of "user-experience designers" has conspired to sabotage the users of their products. I personally think that one of the drivers of the transition from the desktop to the smartphone was precisely the user-experience designers who were turning the "user experience" on the desktop into sheer hell, and the smartphone came out as the simpler version of that "user experience". If email gets so hostile to the user, we the users will switch to texting or chatting, not because we particularly like texting or chatting but because email has become unusable.
You never know what is going to happen to your favorite application when you download an "update". New releases (which you are forced to adopt even if you are perfectly happy with the old release) often result in lengthy detours trying to
figure out how to do things that were trivial in the previous release (and that
have been complicated by the software manufacturer for no other reason than to justify
a new release). A few weeks ago my
computer displayed the message "Updating Skype... Just a moment, we're
improving your Skype experience". How in heaven do they know that this
will improve my Skype experience? Of course they don't. The reason they want
you to move to a new release is different: it will certainly improve THEIR
experience. Whether it will also improve mine and yours is a secondary issue.
At the least, any change in the user interface will make it more difficult to
do the things to which you were accustomed.
It has become customary that a program only gives your two choices: "Download Now" or "Not Now". The option "Never, thanks!" does not exist anymore.
Software engineers have a passion for littering webpages with images and videos, while users simply want to read the text. An image can tell a thousand words but, more generally, in the Internet world of 2019 an image (or, worse, a video) can kill a thousand words because the image or video will take so long to download that the reader may just click on the "back" button of the browser and abandon the page.
As i wait for the New York Times website to finish downloading the article that i would like to read, and i watch the endless sequence of "transferring data from..." displayed at the bottom of my browser, i think nostalgically of the days when i would just flip through the paper pages of the newspaper in a few seconds. Meanwhile, the browser (the latest version from the top Internet company) is still downloading pictures and adverts and doesn't let me scroll down to the few sentences that really interest me. When it's finally done, the search feature doesn't work, probably because the ingenious software engineers of the New York Times used some avantgarde script to challenge the software engineers of the browser.
Incidentally, the main New York Times webpage is structured as a series of graphically appealing blocks. The one in the middle, however, has no picture and only the text: "Error: Server Error. The server encountered a temporary error and could not complete your request. Please try again in 30 seconds." I am not sure which request my browser issued to their server. I just wonder how we would have felt 20 years ago if the newspaper delivered to our house early in the morning had a big note in the middle "Server Error..."
Alan Cooper, a former Microsoft engineer, wrote a hilarious memoir of how software is designed (and unwanted new releases come to be), "The Inmates Are Running the Asylum" (1999), a book that notes how engineers "keep fixing what's not broken until it's broken". Coincidentally, the book came out the same year when Shawn Fanning released the Napster software that enabled pirate copies of copyrighted music and Jon Johansen released the DeCSS software that enabled pirate copies of copyrighted films, two releases that constituted indeed progress... but both inventors ended up in court and almost in jail.
Incidentally, the reason why so many corporations (banks, airlines, news media, etc) want you to update your browser to the latest release is that your old browser does not allow them to track you efficiently, steal your private data, customize advertising, and maximize their profits out of your browsing experience. In fact, it would be much easier for them to deliver their services in a text-only format that even 1990s browsers can manage. They are willing to spend good money on highly-paid software engineers not to improve your browsing experience but to make sure that they can get the most out of it.
We live in an age in which installing a wireless modem can take a whole day and external hard disks get corrupted after a few months "if you use them too often" (as an associate told me at Silicon Valley's most popular computer store).
In the old days i was backing up my work all the time because i didn't trust computers: they frequently crashed. I didn't trust them because they were not reliable. These days computers don't crash anymore but i still back up frequently my work: i don't trust them because it is unpredictable what they will do with my work. Computers download and store files where they want (often in obscure folders/ directories), they change my desktop appearance, sometimes they change the formatting of my documents, etc. The manufacturers that sell them tell me that they are becoming "more and more intelligent": obviously we have wildly different definitions of "intelligence".
In 1997 Steve Jobs famously told Business Week: "People don't know what they want until you show it to them." Maybe. But sometimes the high-tech industry should say: "People don't know what they want until we FORCE it on them and give them no alternative."
Reality check: here is the transcript of a conversation with Comcast's automated customer support:
"If you are currently a Comcast customer, press 1" [I press 1]
"Please enter the ten-digit phone number associated with your account" [I enter my phone number]
"OK Please wait just a moment while i access your account"
"For technical help press 1"
"For billing press 2" [I press 2]
If you are calling regarding important information about Xfinity etc press 1 [I press 2]
"For payments press 1"
"For balance information press 2"
"For payment locations press 3"
"For all other billing questions press 4" [I press 4]
"For questions about your first bill press 1"
"For other billing questions press 3" [I press 3]
"Thank you for calling Comcast. Our office is currently closed."
(You can listen to it at
https://soundcloud.com/scaruffi/comcast-customer-support )
An example of eBay's customer support (Half.com was an eBay service): "Dear Piero, Thank you for contacting Half.com. I understand you wanted to retrieve the items on your wish list. Let me help you out with this concern. Piero, I want to let you know that you can no longer retrieve the listed item on your wish list. I’m glad that I was able to help you. Thank you for choosing Half.com" (04 Sep 2017)
Based on the evidence, it is easier to believe that we still live in the stone age of computer science than to believe that we are about to witness the advent of superhuman intelligence in machines.
It is interesting how different generations react to the stupidity of machines: the old generation that grew up without electronic machines gets extremely upset (because the automated system can complicate things that used to be simple in the old-fashioned manual system), my generation (that grew up with machines) gets somewhat upset (because machines are still so dumb), and the younger generations are progressively less upset, with the youngest ones simply taking for granted that customer support has to be what it is (from lousy to non-existent) and that many things (pretty much all the things that require common sense, expertise, and what we normally call "intelligence")
are virtually impossible for machines.
A book on "The State of Machine Stupidity" instead of "The State of Machine Intelligence" should be much longer.
Incidentally, there are very important fields in which we haven't even achieved the first step of automation: getting rid of paper. Health care, for example, still depends on paper: your medical records are probably stored in old fashioned files, not the files made of zeros and ones but the ones made of cardboard or plastic. We are bombarded daily by news of amazing medical devices and applications that will change the way diseases are prevented, identified and healed, but for the time being we have seen very little progress in simply turning all those paper files into computer files that the patient can access from a regular computer or smartphone and then save, print, email or delete at will.
What we have done so far, and only in some fields, is to replace bits and pieces of human intelligence with rather unintelligent machines that can only understand very simple commands (less than what a two-year old toddler can understand) and can perform very simple tasks.
In the process we are also achieving lower and lower forms of human intelligence, addicted to having technology simplify all sorts of tasks (more about this later). Of course, many people claim the opposite: from the point of view of a lower intelligence, what unintelligent machines do might appear intelligent.
Back to the Table of Contents
Purchase "Intelligence is not Artificial"
| 