Automation is good when it learns. If Google learns not to ask for the cc to your uncle when you email your mom, then its good thing. And if the more you use it and it learns from feedback, that is a better thing.

An expert at something wants assistance with consistently applying her own rules based context to her own work. If the automation learns that context then its OK in my book.

Your considerations are interesting and seem to be about your hacker ethic, which is cool. My no "smart" list is:

- Transparency: Will I know when something "smart" is happening? If I get some feedback when it's happening and ideally can ignore or override the machine's errors, I enjoy the help. Google Maps re-routing options. For this reason, your Gmail suggested recipients example doesn't bother me, because the default is still I'm in control. Most smart lightbulbs and TVs on the other hand just make the feedback loop between press switch and turn on longer and slower, or behave in ways that are hard to understand.

- Human factors: Don't let a "smart" thing do something so unexpected that it causes /me/ to misbehave or I am constantly fighting it. Related to the principle of Least Astonishment. Smart features should be focused on preventing human errors and tested to make sure they don't cause them. It is hard to put your finger on, but a bad example is Microsoft Word's old "It looks like you're writing a letter" or smart bullet points. Whereas Gmail "did you mean to attach a file? You mentioned an attachment" prevents a common post-completion error that humans make.

- Trust: Do I trust a company's manual human review of this "smart" feature and the data they need to make it? Because that's a key part of how they are making sure it's smart. Do they follow good privacy laws and get lots of press scrutiny? No smart features for private notes, private messages, diaries, trade secrets, or always on microphones in my home.

- Security: If something is connected to the internet, who is maintaining its security? Do they have a culture and commitment to my device not getting hacked? Is this device something I am willing to risk getting hacked or is it actually better that it's off the grid?

- Practical: Does it have an off switch? Does it work offline as well as its "non-smart equivalent?

Yes, online suggestions seem so invasive, but it still shows it's imperfection and stupidity. For someone like me, who is an oxymoron - a Japanese born atheist Jew by choice ex-protestant first-generation immigrant living in Harlem with her black family as a jazz composer who plays the violin in synagogues for living, online suggestions only reminds me that I really want a companion and a true friend whom I can talk to and touch, show affection to, in person. AI is not even close to it. Also, I use google translate the languages that I know (some to fluent) and translation are completely unusable and utterly wrong most if the time. (At least for Japanese and Hebrew)
Sorry I digressed...

Interesting thoughts! I agree with most of it, but would personally expand on your #1.

I like building my own automation into things. As a software developer, I agree that I don't want others' rules enforced upon me. But, if there's a rule I think of or a process that is manual and automate-able, then I try to build out my own automation to ease that task. I end up with (what I think is) the best of both worlds - I have automation, and the automation follows the rules that I set!

Speaking of cars,
my brothe-in-law, the neighbourhood mechanic, is sad that now he can't fix as much because of all these new fangled computers.

I'm not an expert driver, but when I drive my car I am not "in the middle of going somewhere," I am driving, enjoying driving, and for this I want no automation.

The only time I ever got an automatic transmission for a gasoline car was when I bought at Christmas and they sold me the floor model, with bells and whistles, for cheaper than a modest version with manual transmission.

Speaking of other people's rules,
My stupid computer has an automatic grammar checker that I often have to overrule, as I know more about grammar than whoever programmed the stupid thing. (Once you know all the rules you know when to break them)

Debian I hope?

Derek, „Linux laptop”? No more OpenBSD? :(
Actually I'm using FreeBSD this year, but thought a mention of that would be too distracting from the point. — Derek

As always, collaboration is about consent.

We each have different needs, at a detailed level, based on our genes and history. So we need to be able to choose the who, what, where, when, and how of our lives.

This is what technology helps us do better, eventually giving us enough options for us to find something that works well for us, uniquely.

Right now we're still in the fetal stages of that technology, and so we have few, or crappy, options mostly.

Isn't it just a matter of levels of abstraction? Sure I can use a rock to beat a nail, but a hammer is more comfortable and a nail gun enables speed.

You choose to use linux, but linux is built on levels of abstraction. Your one command is actually running many commands under the hood.

If AI enables efficiency gains, shouldn't the goal be to learn to use those tools we have access to? AI based spell correct would have helped you catch that "he drive" should have been "he drives" (not trying to point out errors, just trying to give a useful example of advanced tooling being helpful).

Obviously, there is a flip side of enjoyment. Not everything needs to be efficient. Sometimes it can be fun to just put something together with a hammer and the nail gun is too much for the job. Sometimes the UI of a really cool tool is just too off putting to keep using.

I guess I needed to write all of that to come around to agreeing with your three rules. I just lean more towards #3. I want the tools around me to help me as much as possible. And my tolerance for that is higher such that if it degrades some of the enjoyment, that is okay.

I'm also not a fan of most automation though I do like it when my email address is auto-populated as it saves me time and I don't see that as an intrusion. I don't however want my car driving itself, or adjusting my driving as I still want full control in that environment. It probably comes down to (as you noted) how much control you enjoy having over the particular task/adventure.

It seems like we've always looked for ways to create machines to do the things we'd rather not be bothered with. It's just ramped up (exponentially over the last 10-20 years) to a degree that makes many of us uneasy. And getting that genie back in the bottle, if only for a nap, has become yet another unwanted task, another thing we have to become experts at controlling. Does anyone feel like they actually have more free time today than they did 20 years ago? I sure don't. But maybe (probably) I'm doing it wrong.

The older I get, the more feedback I want for my work.

We are all cyborgs, now. Our computers are everywhere and help us in just about every way. This is the middle step between old style computers and computers that will be a physical part of our bodies. Privacy has become antiquated.

I think automation in the context of “personal” may be somewhat concerning but automation in general is something I believe is necessary as we continue to innovate and become a more sophisticated society. We will need to rely on automation in order to make things more efficient. If we were to rely on purely analog tech, sans automation, we’d be a lot further behind in manufacturing and technology. I also think you might be confusing automation and intelligence, given that automation is truly logic based, where as your auto email populate is some sort of machine learning algorithm that fired because you may typically copy your uncle when sending an email to your mother.

I remember having four tv channels with nothing
to watch....

Now I have 1000’s and still the same problem!

Progress..........?

I think automation has a role in our lives. As you pointed out in your blog, it's different for everyone.

For me, I automate everything that I consider a "time sink". Time is the one thing I cannot make more of, so if automation can free up my time to pursue the things that are MOST important to me, then it has improved my life!

Like all things, in moderation, automation is good. ;-)

Jess

Hi Derek,

Sweet and short article. I didn’t think much about automation until I read your take on this. I do enjoy some of the automation plugged in my cel phone for convenience. But the one thing I find annoying is when I get a message from google maps telling me how long will it take to get to the next destination without me telling google maps where I want to do. I think I’m general we are very predictable, we like to do and go to the same places, which is an advantage for any sfw that is wired to predict pretty much your next move. My main concern with automation is the fact that it makes life easier in many ways that you might not need to think hard anymore, therefore, using your brain less and less. Thanks for sharing your thoughts on this topic!

I value automation if it is better at doing the tasks than I am. Mostly these are less creative, more mundane actions. If a machine is more consistent than I am as a forgetful human, then I am open to trying it out.

BUT

I must understand how it works. If I don't understand how it works, then I am not comfortable subscribing to the automation. I think this is the intimidating parts of machine learning. What is behind the magic of these recommendation engines? If I can understand it, then I will move on to considering it. If I don't understand the fundementals of how it works, I stay away...until I maybe understand how it works.

went to an interesting book presentation more or less about this:
"la fin de l'individu" by Gaspard Koenig. french philosopher investigating how AI touches our free will (i'm still reading the book).

oh...sorry i know quoting is wrong :) forgive me this time.

thanks for the blog Derek.

I've good with some automation -- I don't mind suggesting how to finish sentences in template-type emails, or even suggesting email addresses.

But, other automations I prefer to keep manual. Like sticking to hand-writing first drafts of blog posts and books instead of typing, using record-then-transcribe or using writing software. Too much of the shitty first draft is shitty and it needs to be on paper, sketched out first before I begin to organize it for reading.

Expert is one way to put it. Although I see it more as people 'set in their ways'. When someone does something, on a regular basis with purposeful intent, they tend to get 'set in their ways'. They become clear about their boundaries of authority and comfort.

With the number of hours you've spent on a computer, learning how to code, building companies, writing books - you have a certain relationship with the machine.

Have you read I, Robot by Issac Asimov?

I believe the arbiter in this question is: does achieving the desired outcome require a novel solution?

If not, and especially when others' lives are being staked, you're probably better off letting the robot do it.

the thing i would find super creepy is if AI gets to master emotional intelligence and where we cry on the robots shoulder where imitation becomes indistinguishable from real heartfelt issues too creepy to contemplate
we need to deliberately opt for less dont we?

Note sure if my preference for automation can be put into a formula yet.

I notice that I prefer automation both in cases of things I do not care about (like you mention cars) & for things I care about more. I do like auto spell correction to ensure my recipients don't have to go through gibberish messages. At the same time, I'd rather not spend the time / effort for the corrections myself.

I think it comes down to 3 things, and you've mentioned 2 of them.

The first one is being able of having your own rules. This is important because some undesired rules may lead to undesired outcomes. If you need to stay on top of the automation because it's unreliable, it defeats the purpose of automation.

The second is if you enjoy doing the task. I usually like to cook, and not only for health reasons. I could buy prepackaged food or order from restaurants, but I enjoy the process of cooking so I often do it myself.

And finally, understanding. I think that's were the creepiness you mentioned comes from. If it felt creepy, it's because you didn't know how they did it. It felt invasive because you thought (you were probably right) they were analyzing your previous conversations and personal data. But, what if they were doing it in a privacy respective manner? What if they had some JavaScript running on the client and looked for associated emails? Then it wouldn't be creepy nor invasive. Maybe you still wouldn't want it, but then it'd be related with one of the previous points. I think the key takeaway here is that we don't want to automate anything that we don't understand, it's only once we are doing something tedious and repetitive that we want to automate.

Think I am going to enjoy your dailies

...or, if you use those technologies to free you up to pursue your dreams. For example, I left a high-tech environment (for it's time) in 1995 and slimed down my tech needs by trading it all for an acoustic guitar, paper, and pencil, laptop and printer and began composing in earnest. These tools and some recording and manusçript software still meet my needs today. I am free to create and I can record and refine by blending old and new technologies.

I agree completely. I hate that we are subject to buying and selling of our information. It is so pervasive and I wish to remain as anonymous as possible. I don't have a facebook, or Instagram account contrary to 90 plus % of people out there. I watched ex-machina and am amazed that, this is going to sound cliché but, a carbon based life form that is trying to replace itself obsessively by a silicon based artificial intelligence. I think we are pushing the boundaries of technology a little too fast. I am old fashioned in the regard of having patience when it comes to important aspects of living a simple straight forward kind of life.

I work for a semiconductor manufacturer and my career has always revolved around software/computers in some form or fashion - consequently I am very distrustful of automation and computers for some things. One of mine is cars. Self driving cars terrify me. I hate the proliferation of computers into cars and automotive systems, because it's a black box that does mysterious unknown things that I can't always understand. Whereas a simple on/off switch - that's easy. A computer version of an on/off switch has many, many points of failure and can even get sidetracked by something totally unrelated to the on/off function. On the other hand, I love what technology and automation has made possible for musicians. So, I guess I'm human and inconsistent.

Tim Ewald argues that we should value "hand tools" as opposed to power tools because automating a task changes the way we think about said task. Really great talk, especially if you have any passing interest in woodworking and/or programming: https://www.youtube.com/watch?v=ShEez0JkOFw

Your think 🤔 does not make sense about we are human as birds as biotechnologily

I would add that it's also good to have automation for things that do matter a lot - like high stakes situations where the problem is very complex or precise (e.g., surgery, diagnosing illnesses, predicting the weather, etc.). Of course no automation system should truly operate unsupervised, for everyone's sake haha

Derek,
1) Is technology at its best when it imitates nature in its manner of operation?
2) Any sufficiently advanced technology is indistinguishable from magic.
3) At what point are we no longer able to ascertain the boundary between ourselves and the technologies we are using?

Took a moment to share my thoughts on the subject in a separate post.
https://manuelmoreale.com/smart-technology

“…because teaching people how to use technology properly would be way harder.” = Great point! — Derek

For me, I feel automation, AI and the likes, deprive me of intellectual pursuit and that comes from a guy who worked in the field of machine learning.

I think it should be heavily used when it comes to taxpayer funded services because we seek efficiency, nudge, autopilot and a more or less tailored solution for everyone.

It should also be used when people's life are at stake: transportations, factories, etc.

But when it comes to personal spheres, both approaches should be available. On the rare occasions that I browse social media, I want my feeds to be linear not curated.

When I am reading a book, I sometime obsess about details, a very small idea, a neat analogy. If I read a computer generated resume I don't find any joy in it. It does not nurture thoughts.

I've been contemplating things like this and it seems that for me it comes down to #3 (I want the assistance) in your list. Often times (particularly for work email) I love when it suggests others I might like to include because I don't have to think about it and it was helpful to me. Auto correct is beautiful when it works but if every time I write a word it changes it and I have to proof read and change it back, it should stop changing it.

If there is a reason for ME that you are tracking if I open an email or a message (say to mark it as read on all my systems) then that is valuable to me. (But if it is for you or someone else it feels invasive.)

I often don't mind ads based on what I am looking at now, something ideally that could help me make a better decision. But don't push me down a path of things you think I want.

I particularly love automation I can opt into and setup. Can my smart speaker turn the volume down at a certain time so if I ask a question in the morning it doesn't wake my kids? Can I have my lights turn on at dusk, or my to-do list to pop up when I get to the store?

But as soon as someone is frequently pushing their own agenda at me through whatever mechanism (spam, robo calls, irrelevant clickbait articles or ads online and even on the radio or tv) I start to feel a resistance where I want to push back. I think this has always been true for me. I remember back in college we started muting the tv when commercials came on because they were louder and annoying and irrelevant.

MIT Technology Review:
China has started a grand experiment in AI education. It could reshape how the world learns.
In recent years, the country has rushed to pursue “intelligent education.” Now its billion-dollar ed-tech companies are planning to export their vision overseas.
by Karen Hao
https://www.technologyreview.com/s/614057/china-squirrel-has-started-a-grand-experiment-in-ai-education-it-could-reshape-how-the/

I see some concern with "being used" and privacy.

My local university professor is telephoned by journalists from all over North America as he's a computer expert. He has a very good book that I attended his one-day seminar about, called

Technocreep: The Surrender of Privacy and the Capitalization of Intimacy

The book is by Thomas P. Keenan

Automation in public restrooms has been a pet peeve of mine for years. I am a software engineer, so in a way I automate things for a living, but among the many things that drive me crazy are:

- Faucets that don't turn on when I put my hands under them
- Faucets that still haven't turned off by the time I leave
- Paper towel dispensers that treat me like a ghost
- Toilets that decide to flush themselves while I'm still using them
- Soap dispensers that like to squirt soap on my forearm and/or shirt sleeve because they were mounted on the side of the sink

Why didn't I get to vote on this? Am I the only one who appreciates the simplicity and freedom of flushing my own toilet, or am I the only one that can handle the responsibility of doing so?

Well, I like that you are one of the few people (including me) who try to stay away from technology even though it seems impossible in today's world, but in my opinion, I think you somehow contradicted yourself in this article. You said you don't mind if your phone does these things but you do mind if your computer does similar things.
Actually, phones these days are even smarter than our laptops. If you really think about it, phones are just smaller version of laptops because you can practically do everything on a smart phone. What's more, laptops only recognize what we type in & gives us suggestions that may or may not be helpful to us. On the other hand, phones store our data as well as our VOICE. More often than not, Google will give you suggestions about the things you had been talking about a few hours ago, even though you never particularly typed it on your phone ever! What's worse, we carry our phones everywhere & not our laptops.
Now if you're NOT using a smartphone, that's a different story.
In conclusion, what I would say is that our computers are much safer than our phones. Google(Gmail) is one of the few companies that gives you suggestions to make your life easier, and also gives you options to turn off that feature completely. If you wish, I can send you the directions to turn off suggestions.

I'm glad you wrote this article, gave me a chance to really think about this topic!

Regards,
Lehar
The contradiction was kind of the point of the article. Being OK with AI in some places and not others. I don't use my phone much, and I don't use Google ever. Yes it's a powerful device but I don't use it for anything that matters much to me. Whereas I use my computer most waking hours for creating things that mean everything to me. — Derek

Massively agree, and I also think there's an interesting chicken/egg thing here, too. Dropping automation can sometimes lift the veil for people who don't score highly on your three conditions.

I've been happily writing code in modern IDEs for a couple of years*, but this month I switched to Vim. Thing is, I'm absolutely not an expert at what I do, and development really is just a job for me. Until now, I just used an IDE because most other web developers do.

But then I realised VSCode / Sublime etc aren't Lindy, and figured it was a safer long term bet to switch to Vim.
https://medium.com/incerto/an-expert-called-lindy-fdb30f146eaf

The unexpected consequence is that I'm enjoying my work a _lot_ more. It's such a nice feeling when fuzzy concepts get clearer, and when there's less unknown stuff between me and the machine. But I had no idea that my attitude would change when I started learning - if anything, talk of tech purism was a turn-off for me beforehand.

Switching has actually pushed me further along on each of your three factors! Increased expertise, increased enjoyment and removed desire for tools. Makes me want to experiment with other things along these lines.


(*My career change was a direct result of reading your '/prog' and '/learn-js' blog posts! 👍)

Hello Derek, if ever you were to go to a song on Youtube, this is the one: (But I'm linking it here for your readers, not for you)

Jason at #12 said we are all cyborgs now, etc.

Indeed, and here's a song about it, made as an assignment for a night school class. 'Tis a whimsical parody of the first song ever played on MTV.

https://www.youtube.com/watch?v=RO0-7YAxxDY&feature=c4-overview&list=UUAiABuhVSMZJMqyv4Ur5XqA

Both Richard Feynman and my old boss Sir Terry Matthews made the point that they used to fix radios for elderly neighbours back in the day (usually by swapping two of the three vacuum tubes) while today radio internals are miniaturised on electronic circuits. The MIT particle accelerator internals were all sealed away with a nice UI, while the Harvard one was a mess with cables everywhere, it was always breaking down and fixed back up with "duct tape".

Apparently the Harvard one is where all the meaningful discoveries were made.