Please don't send me smart replies
engadget.com
engadget.com
Are your friends selfishly taking up time trying to talk to you when you could be adding more people to your friends list instead?
Do you have a craving for increased amounts of non-genuine social interaction?
Has your life become so depressingly repeatable that our algorithms can accurately represent your superficial chatter?
If so, we've got something for you: SMART REPLIES!
SMART REPLIES! For when your communication is so valuable and necessary it doesn't need you!
SMART REPLIES! Yum clams! So cute!
SMART REPLIES!...if you thought twitter and social-networking was inane, you ain't seen nothing yet...
Edit: I apologise, i just had a reflex reaction. Wow...this is the world we're making for ourselves...
Should we save our outrage for something more then a feature in an announced chat app?
..and still feel a little bit guilty for being so bloody shallow.
I'm sure this will be downvoted (edit: maybe I was wrong, which makes me feel better yet) because it doesn't really add anything to the discussion, but I've been increasingly depressed about this stuff lately, and hearing/seeing someone else basically put the words in my brain down here, helped a lot.
Being surrounded by people who think that anything involving computers or technology is automatically good and something to be supported and defended no matter the cost to our freedom, humanity or privacy, is incredibly frustrating. As someone who first pushed many of these people towards the internet and technology in general, I almost feel responsible for what they've become.
As someone who has done very well as far as careers go, thanks to an early start with computers and technology as a child in the late 70s / early 80s, I know without a doubt that I'm partially responsible for some of what I hate now, whether directly by way of cash-induced moral ambiguity, or indirectly thanks to projects I've contributed to.
I love technology and progress. I just wish that progress actually meant doing important things, rather than finding newer, more invasive ways to surveil and manipulate people as it has for the past decade or more.
Mobile phones may sound like a good technology until you realise that people are using them because all of their friends are miles away, or because they have to and their boss calls them all the time, or they are so bored at work that they would rather browse facebook.
Cars are something that could allow you to travel long distances but we mostly use them to sit in traffic or travel long distances simply because we have paved over everything around us and nature is so far away. Cars have contributed to pollution and cities that resemble car parks.
To me, and this is obvious from my experience in the jungle, the modern world is a complete failure that has not improved our lives at all.
My best experience in life was away from the modern world. I mean my mental and physical health ended up being completely different by the third week. My experience of life.... I was actually living. I have never felt that since. I feel like I know a secret about the world that others who have never had that experience will ever know.
I have a feeling that modern technology and life is guided by corruption, not by what is best for humanity.
That's only one step away from justifying destroying the rest of our planet because we can just "experience" it with technology anyway.
What's funny is that I'm sure there are people who make lots of money working for Facebook, Google, and others reading this thinking "we're not evil, our slogan even said so! This person is nuts!", but nobody really thinks what they're doing is wrong while they're doing it -- or they wouldn't be. Hindsight is 20/20 and all that..
If someone sends me a picture of their cute dog, I'll happily acknowledge it. But I don't see the difference between picking a somewhat simple, canned response over writing essentially that same thing out myself. It's about acknowledging that you received their picture and agreeing that it is cute.
There are many things where automating stuff feels wrong. But most of these involve some kind of manipulation: help with writing a sweet message to your lover, optimizing your email response to the particular personality of the person you're talking to so that they'll 'like it more'. Stuff like that.
But automating away things that are kind of inane to begin with doesn't seem like such a big deal to me. It's no different from the 'hello how are you' and 'how was your weekend' talk. It's not quite pointless (in my view), but it's also not something so authentic that any protocol or automation in the matter is a bad thing (in fact, we already do this).
Letters to friends, or letters with significance (e.g. a resignation letter) will of course always be written by hand, word for word, but there's room for smart replies to be useful without having it diminish the human race
> What are you up to?
>> Work.
or
> Want to hang out?
>> Sure.
The suggestions Google provides are a little bit more upbeat and amusing to me.
> What are you up to?
>> I am working!
or
> Want to hang out?
>> I'd love to.
I especially like how it inverts how I'd normally use exclamation marks and periods. "I'm working." "I'd love to!" etc.
Anyway, neat feature. The author doesn't like it, I do, so she can not use the feature and I can use it :P
I don't think anyone is going to think "I'm not really sure if I love my partner, but Google says I do, so I guess I do." I also don't think people are using mobile apps to have the deep types of conversations where this would matter.
Like I said, I see it as a more interesting spellcheck. I usually say things like "yup", Google offers more amusing suggestions that I enjoy.
(Looking forward to the day that that sort of reply can be generated. I'm sure it would be far more fun than "wanna hang?" "ya")
Sometimes I think AI isn't advancing at all, its just that we're all working more and more to fall in line like good automatons...
I love sarcasm, dead-pan, violently black-humour...and I'm watching it all slowly disappear and we all become pidgeonholed, public, on-show, quantised and sanitised.
I'd much rather bring back:
>What are you up to? >> In middle of school shooting. Cops on way. Call you back in 5?
>Want to hang out? >> I'm in the bathroom, so in a way, I already am...
> sends through that picture of child holding flower >> Sometimes that shark, he looks right into you. Right into your eyes. You know the thing about a shark, he's got... lifeless eyes, black eyes, like a doll's eye. When he comes at ya, doesn't seem to be livin'. Until he bites ya and those black eyes roll over white. And then, ah...
Oh, but the two are unrelated you say...
Well no. I'm not certain that they are...
These days, that sounds like something that would make SWAT invite themselves into your house, through a window.
The responses are sure cute and creative. I seriously doubt whoever writes like this today will change their ways because of "smart replies". Most people don't write stuff like that, definitely not during most of the conversations.
I don't really have a formed view yet. I think getting outraged over someone using augmented software to respond faster is wrong, and only a problem with the person who gets outraged. On the other hand I do appreciate the effect of limiting the potential conversation space if people get too dependent those reply suggestions.
It is also true though that a lot of conversations one has over e-mail or IMs follow so common patterns that there were many times I wished I could just automate it away, to have a bot reply to them without the other side knowing. Maybe I'm just antisocial? But I tend to differentiate between requests for information from requests for personal attention, and wouldn't mind having my bot handle all of the former.
(Then again, I wouldn't accept anything except a bot I wrote myself, so maybe I'm not that antisocial - since I'd be just spawning a part of myself as an external process. That's also one reason why I'm so pissed about all the modern bot / AI tech from Google, et al., which has zero configurability, zero ways to plug your own stuff in yourself.)
This seems similar or related to the need people seem to have to remain permanently connected even to people who they only met for a few seconds at a bar once, to increase that ever-important "friend count". Some relationships are meant to end. It's not healthy to try to maintain 1000 vaguely familiar people as "friends".
The ridiculous thing is, most of those "friends" are dreading having to pretend to care about the things you, the guy they met for a few seconds once but feel bad deleting, post on the internet as well. Just stop it. Neither side wins in forced out-of-sympathy/fear/loyalty/obligation/etc relationships.
(Disclaimer: I don't use allo so I don't know what it's capable of. I kind of doubt it's there yet. )
It's not. It's bigger and faster, and being applied to situations it wasn't previously applied to.
Conversions will become more predictable and shallow then they already are.
> Want to Hang out?
>> I'd love to!
>> Nah, I am busy.
In that case, I don't think anyone should have a problem.
The simulated conversation image is a great example. After seeing those responses, the response 'aww so cute' or 'Love the daisy' are about about the same or better than what I can or would care to think of in a text conversation. Is there really a better way to think answer that and why should I waste the time to do so?
Do people on chat with their nearest and dearest? Or do they chat with a mix of people who they may like or even dislike.
It also depends on the discussion. Whenever someone talks to me about something requiring my personal attention, I pay that attention. But most conversations with people are trivial, and expecting someone to drop everything they were concentrating on and pay full attention to your trivial question is, again, a sign of disrespect.
This effect is the only one thing that worries me with those replies. It won't be good if people start to choose only from suggestions instead of choosing them only when they are appropriate. And, as you observed, the effect can be subtle.
But the very idea of a computer typing up the response for you? I don't mind that at all - that's what computers are for, to let you do more with less effort.
---
With Smart Replies (as with other communication-helping technology), meaning comes from how and when you choose to use them, and there's no reason you shouldn't choose to use them when it communicates what you were intending. If I was going to say "Yes!" and that's a Smart Reply option, why shouldn't I save the time and choose it? It was still my action that sent that message, still my choice to send it, still the meaning that I wanted to communicate. Smart Replies save time for the simple stuff, so you can better spend your time on other messages that you might write manually.
Text messages, twitter and snapchat may have made conversations lighter, less substantial and more throwaway, but they have cultivated a rich flora of memes, silly jokes, new ways of sharing. Even emoji, which at first blush also seem like a very lazy way of communicating, have taken on cultural relevance and can be used in interesting and fresh ways.
A smart reply is a smart reply. The best case scenario is that it was close enough to what you were going to say anyway that the small difference in content is outweighed by the reduced effort of not having to type or think. It cannot lead to new forms of creative expression because its very use demands not expressing yourself more succinctly but instead expressing yourself less precisely. Expressing less of yourself.
>Further, I don't care how intelligent these Smart Replies are: They can never capture the personality and character of a real human conversation.
Really? Tell that to the people living in the time before Alphago made it's 37th move and astounded Lee Sedol himself. AI is here, it is as smart or smarter in some areas. The number of such areas is going to grow. Get used to it.
(And while we are on the subject.. There was a blog post back in 2005 about the inverse situation: if the person behind the other screen has the preconception that you are a bot, convincing them of your humanity is nearly impossible. The post now lives as a PDF document.[0])
0: https://msu.edu/~pennock5/courses/ALife/Striegel_Failed_Turi...
This and indoctrination of presence of surveillance is the problem with all those features.
[Edit: spelling, clarification]
Then again I remember seeing on the palm pre a feature that showed you notes about the caller on the call screen. Supposed to be for sales people to put stuff like "Wife Jane, 3 kids, Bobby, Sue, Jamal" to read before they picked up.
Is there anything wrong with that? Seems very useful.
(How do people think how previous generations handled having many acquaintances and still knowing something personal about each of them? They noted things down, on paper.)
Anyway he would go on dates about 3-5 times a week with new people. He also ran into a lot of crazy.
He's always been more of a surface level relationship person. We were talking about this last night and he's just coming to grips with having a real relationship that has settled and become comfortable as opposed to awkward. He's somewhat at a loss. I'm telling him that now is when he can really have fun. But that part is anecdoteal I guess.
Text messaging is a lazy way to have a conversation; this is a perfect fit. If you want to have a real conversation, call me, or meet me for lunch. Or do the former to arrange the latter (novel!).
It doesn't mean e-mail isn't perfectly fine for a lot of conversation, but you should expect people to treat it accordingly: You care enough to write an e-mail instead of an IM, but not enough to call them or a arrange a face to face.
Sometimes it's perfectly appropriate to convey that. Sometimes it's not.
However, it is funny to see "normal people" getting more and more used to this kind of living while slowly more and more tech people find it disgusting. Imagine a future where only nerds are capable of small talk!
I'd reckon smart replies would follow a similar path. Might look odd, but would become natural in a few years.
I think the problem lies with the examples, and frankly, it is up to you how you use the feature.
Yes, if I write my SO "I love you", I expect him/her to reply with a honest "I love you too" that was not actually a computer replying. It is a philosophical question obviously, whether you see a response written by somebody or by the algorithm, if he/she actually selected it.
But if I just get a "I'm 5 mins late" from my friend, I don't think a "Don't worry, I'll wait" suggested by the algorithm is that bad and I myself would not mind getting such a "algorithmically suggested" response by my friend. After all, the other party selected the response himself, so I think it is honest.
I'm pretty sure Google had these or similar discussions internally as well.
Except, suggesting "Don't worry I wait" means that the person might click that, instead of typing "ok".
This is a problem if you know the person well, and know that a response of "ok" in the situation indicates they are annoyed at you, etc.
Maybe I did not understand you though?
> I can understand canned responses like "Be right there!" or "I'll be late!" if you're in a car and only have a few seconds to reply before you have to get back to the business of driving
I can't. Please do not use your phone while driving. You are not just putting yourself, but others around you in danger.
See, I think having ASCII completely misses the point of a blog post. I can understand using canned characters for article indexes so that a publication can have a consistent style. But for the article itself, then ASCII seems like an unnecessarily lazy way to write a post.
In this blog post, for example, Nicole gave a demonstration on how you could use italics to stress particular words or "--" to emulate an em-dash. Are you so devoid of creativity that you can't stylize your text by yourself? Is writing so much of a hassle that you can't use the appropriate characters before posting?
Further, I don't care how readable these ASCII posts are: They can never capture the personality and character of a hand written letter. Having the option of ASCII encourages me to give a more straightforward response rather than coming up with something that's perhaps more emblematic of my real personality. It's dehumanizing.
But more than that, using ASCII instead of your own lines and curves robs your friend of you. It means that you don't value the friendship enough to fully engage in the conversation. You'd rather have a robot do all the rendering for you instead of spending time and energy on them. That's terrible.
I know, I know, ASCII is optional. You don't have to use it. But if we're any kind of friends at all, I sure as hell hope you don't.
It will only work under very generic situation with borderline functional replies. For me, that kind of conversation is just a small part of the every communication.
Human conversation is more organic than the demo tries to lead you to think. Details in the words and sentences convey very subtle but identifiable personality traits that the current algorithm cannot capture.
So, it is a really cool feature and a great showcase of recent advancements in applying deep learning to natural language. But it is pretty much a gimmick for now, and I will stay sceptical about whether this is going to turn user around and become a killer feature of Allo, after all if my friends didn't use this, neither will I.
The "smart" replies seems to robotic and i'm sure many people would find it rude if they got them.
- always pressing the first option
- randomly choosing an option with their eyes closed
- pressing on the microphone and speaking gibberish or making gutteral sounds
I'd watch that
I think smart replies are a great advance towards communicating quickly and efficiently - which is the use case for most text conversations. But it is important to realize when we expect communication to be more descriptive and personal.
> I can understand canned responses like "Be right there!" or "I'll be late!" if you're in a car and only have a few seconds to reply before you have to get back to the business of driving.
I typically only use them when I'm busy and I see a message and respond to it from my watch.
The idea of automation is automating things we don't want to do, talking to friends is definitely something I want to do myself. What will be the next thing, a robot that has sex in our place?
OTOH I'm curious as to where this will lead in the future... smart replies replying to smart replies and generating an entire conversation between two people without them even reading it? Telling your robot "Hey, robot, I want you to find me a guy between ages 25 and 30 for a date tonight, he should be dark-haired and like dogs" and the robot interacting with other robots until it finds someone compatible? That sure would be handy for some shy people :)
I prefer end-to-end encryption to yummy clams.
This is another example of the emptyness it will give.
~Mark Manson