Slashdot Mirror
Gmail Smart Replies and the Ever-Growing Pressure to Email Like a Machine (newyorker.com)
An anonymous reader shares an article: I don't use the phrase "Will do!" much in daily conversation, but lately it has been creeping into more and more of my e-mails. An editor asks me to get a draft back to her tomorrow? Will do! A friend heading back to Los Angeles from New York sends me a quick note telling me to enjoy living in the "best city in the world." Will do! The hosts of a panel I'm moderating need me to send over a three-line bio? Will do! "Will do!" is just one of many Smart Replies that Google now provides as a default feature in Gmail, there to assist you in your message composition unless you choose to manually turn them off. In October, the e-mail service, which one analytics firm suggests hosts about a quarter of all the e-mails sent worldwide, made this feature standard on its 1.4 billion active accounts, along with a menu of other innovations.
These include Smart Compose, a feature that finishes your sentences for you with the help of robot intelligence, and Nudges, a feature that bumps unanswered e-mails to the top of your in-box, making you feel increasingly guilty with every sign-in. As with many technological updates that are suddenly imposed on unsuspecting users, the new Gmail interface has been met with much annoyance. When my in-box started offering me Smart Replies, I felt a little offended. How dare it guess what I want to say, I thought. I -- a professional writer! -- have more to offer than just "Got it!" or "Love it!" or "Thanks for letting me know!" (Smart Replies are big on exclamation points.) I started to resent the A.I., which seemed to be learning my speech patterns faster than I could outsmart it. Just as I decided that I'd thwart the machine mind by answering my messages with "Cool!", the service started offering me several "Cool" varietals. Suddenly, I could answer with "Sounds cool" or "Cool, thanks" or the dreaded "Cool, I'll check it out!"
These include Smart Compose, a feature that finishes your sentences for you with the help of robot intelligence, and Nudges, a feature that bumps unanswered e-mails to the top of your in-box, making you feel increasingly guilty with every sign-in. As with many technological updates that are suddenly imposed on unsuspecting users, the new Gmail interface has been met with much annoyance. When my in-box started offering me Smart Replies, I felt a little offended. How dare it guess what I want to say, I thought. I -- a professional writer! -- have more to offer than just "Got it!" or "Love it!" or "Thanks for letting me know!" (Smart Replies are big on exclamation points.) I started to resent the A.I., which seemed to be learning my speech patterns faster than I could outsmart it. Just as I decided that I'd thwart the machine mind by answering my messages with "Cool!", the service started offering me several "Cool" varietals. Suddenly, I could answer with "Sounds cool" or "Cool, thanks" or the dreaded "Cool, I'll check it out!"
"If it is what you say it is, I love it!" - Prisoner 85-3824-AA
I try and stay away from the so-called "smart" replies. It feels insincere to have an algorithm write my response for me.
The things you miss when you actually have an email client. :)
I'm ok with canned responses, but I wish they'd drop the damn exclamation mark from every freaking option. I don't scream "THANK YOU!!!!!" or "WILL DO!!!!!" when I talk. Every damn option they give me has an exclamation mark.
Oh, while I'm bitching, they should also add a newline break after their responses. It looks rather dumb without a line of white space between your signature and their canned reply. I've been secretly hoping that GOOG machine learning notices I remove the ! from every reply and add that line of white space myself...this is machine learning...right?
You don't have to use Gmail. You can get your own hosting with your own domain and many GB of storage for a little as $1 a month!
Hell, get a fixed IP address, and run your own server at home. (Ask a friend if you don't know how to do it.)
Besides: You don't even have to use e-mail. It's for old people anyway. ;)
If there is a pressure, it's to actually treat a computer like a computer!
In other words: To use automation! And make things easily automatable too!
E.g. not act as if one IP address (or one user name, or computer, or cookie, etc) equals one person.
Or force people to use shitty point-and-click interfaces that don't offer any equivalent of loops over lists or of SQL.
If you run an auction site, you have to expect people bidding in under one microsecond, and sending millions of bids per second. And you also have to expect that, and be able to use it, if you use such a site!
(That's why any kind of online review and vote service is bullshit, and will never be reliable or trustworthy. Ever.)
... why can't it do the rest of the job too?
Me, I do my thing and expect the computer to do other things. Mix the two at your own peril.
What would they prefer? *Fuck you*?
“He’s not deformed, he’s just drunk!”
Me too.
Imagine being such an NPC, that you can be replaced by a trivial script. Embarrassing.
Suck my
Blow it out your
Fuck you and the
Make America
Trump is a
Hillary is a
Let me know.
The dreaded "k."
The danger from computers is not that they will eventually get as smart as men, but that we will meanwhile agree to meet them halfway. (Bernard Avishai)
I'll wait until the 'Done' version.
How about a "smart pickup line" feature I can use at the singles bar tonight?
I've abandoned my search for truth; now I'm just looking for some useful delusions.
No, sorry Google, I'll write my own fucking emails, in my own writing style, with words of my choosing ... I'm not interested in your bullshit predictive algorithms trying to inject themselves into my communications.
This is literally one of the dumbest and annoying features I've ever fucking seen in a while.
Sorry, but I can type fast enough and have a good enough grasp of English that I don't need your goddamned fucking help.
Google really are becoming assholes.
I think this is what lead to the borg. The Hive Mind of auto-complete.
Personally I think an AI is smart enough to have suggested phrases in your reply, you really shouldn't be saying this over email at all, or just plain shouldn't be saying it.
In other words, automated response convey so little information, why say it in the first place?
Composition suggestions, piss off.
This is where having a mail client, be it Thunderbird, mail.app, or even Outlook comes in handy. This completely bypasses these types of shenanigans and psychological tricks.
Yes. Oh, how much I will agonize over being able to type standard replies faster when I'm typing them. And angstily bemoan the demise of our society if I have to notice the lookahead standard reply while I am typing my own. /end sarcasm.
Seriously, this is EXACTLY the kind of thing that's great about computing. It lets humans save time on repetitive tasks that our brains can handle more quickly than we can handle with current tools. This frees up time for EVERYTHING ELSE in life. If you're complaining about this, you should strive to one day graduate to the point where you work on fixing #firstworldproblems.
I can't believe people who should know better are still using GMail. Real email is literally $1/month.
I don't respond to AC's.
It is all just words in a page as far as I am concerned
Was your post a Gmail smart reply?
I'll see your senator, and I'll raise you two judges.
Lord Refa: Ink on a page!
I'll see your senator, and I'll raise you two judges.
How do I set my default reply to "Up yours!"
Have gnu, will travel.
Just what I needed: another reason to avoid gmail like the plague. It's the email account I use when I fear that I might receive spam from a new contact or a vendor (that wants me to set up an account to receive "special offers"--which are almost certainly not that special). I wonder if Samsung or Comcast will be impressed with replies to their special offers littered with snazzy "Cool!!!" exclamations?
If this is what being a near-monopoly thinks is innovation, bring on the anti-trust legal eagles.
CUR ALLOC 20195.....5804M
I don't use their stupid canned replies for exactly this reason.
Just because you are weak willed enough to modify the way you communicate just because gmail tells you to doesn't mean the rest of us are.
and look over the cliff, you'll see the packaged on: and best before: dates.
Where are we going and why are we in a handbasket?
Just tried it though, and google composition suggestions is limited to the most bland cliches imaginable.
It will take a lot to get a poetic masterpiece out of it.
Where are we going and why are we in a handbasket?
Rather saddens me, actually. I'm trying to remember the last time the google had a good idea without 10 bad ideas piled on top. Email is an especially sore wound, since there' so much room for improvement there.
I've given up wondering how the google profits from supporting scammers and spammers. Makes as much sense (= zero) as wondering why they don't fix the moderation on Slashdot.
The specific email feature I actually want the most right now is an email system that will bounce any confidential-mode email that anyone tries to send me.
Excuse me, but I have to get back to grading my student's homework in email. I know that this new quick-response is supposed to be time-saving rather than dehumanizing, but it certainly is useless to me and I think I will pledge NEVER to use it. Just like the confidential-mode BS.
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
I don't even like autocorrect turned on, I sure as fuck don't want the gods-be-damned computer to finish sentences for me.
If you can respond with, "will do," then you should't respond at all. I have enough e-mails to read without your meaningless response. If a robot can say it for you please, for the love of all that's holy, just don't respond at all. I won't be offended. If we continue this way the next killer feature of e-mail services needs to be recognizing and ignoring e-mails that other services automatically write.
Those gateway pages that appear when you connect, where you have to say "I agree", are easy to automate away. (I haven't seen one with a CAPTCHA yet.)
You could run a server 24/7 on them, with ever-changing MAC addresses.
I wonder what it would take, for them to bring out the radio interference truck, locate your device, and shut it down... ^^
Can you set it up to say "won't do?" If so I can save a lot of time writing e-mails to my boss.
Just this week I was pondering the same thing with the choices offered by Google's SMS message App on my Android phone.
I came to the conclusion we are training Google's natural language AIs to understand the message we are responding to. The response we make allows the algorithm to categories the original message. If we type something new, we've just told the AI it got the choices wrong.
Actually choosing the dumb responses is confirming the AI natural language understanding, the positive feedback stimuli, typing something new is training stimuli.