Can M wade through phone support menus and cancel Comcast subscriptions? I look forward to a darkly-humorous future in which we pit poorly-paid third-world citizens against each other in wars of call center attrition. Better, once we equip them with those sound-board UIs that play pre-recorded answers in native English-speaking voices (can't find the link), English can become a transmission protocol that few people deal with directly.
In Greg Egan's Permutation City, they have interactive 3D video email, and interactive 3D video spam, and interactive 3D video spam filters.
The spam tries to act like a perfectly normal message as long as it is talking to the spam filter, and as soon as it thinks it is talking to a real person, it shows its spam message. The spam filter tries to impersonate the recipient as best as it can (in 3D video), meanwhile trying to figure out whether the message is spam.
The spam filters are unfortunately humstrung by the fact that they can only become close to real conscious AI and not further, because taking them all the way there would mean you'd be exposing a conscious being to spam all its life, which would be torture and thus criminal. Spammers don't care.
IIRC, this is just a side anecdote in some paragraphs somewhere, but I love it.
I just finished this book. I loved it. Any other recommendations for books like this? I read Accelerando, and am reading Chasm City now (both highly recommended).
I loved _Diaspora_ as well. And the rest of his books, but those two stand out.
I don't know anybody like Egan, he takes the science part to such extremes... My favourite SF author is Iain M Banks, but you're probably already familiar with him.
I would hope in the future that more and more services are actually directly available to be interacted with over one or more messaging networks, enabling easy, asynchronous communication.
Whether Comcast would want to make it that easy for people to cancel is entirely a different issue though :)
The implication to me is that the chat is with a human, who is using an AI tool with the intention of training that tool. What better way to train a new service than to launch it, then answer all the weird, unexpected questions with humans? Gradually more of the questions get answered, the AI gets better trained, and the human-AI becomes increasing more AI.
Further, as the AI gets better, the human working with it has to do less, so they can roll out the service to more users without requiring more staff. Perhaps eventually, no human is needed.
Any good chatterbot has a script or set of scripts to answer questions about its nature, relationship to its creator, etc.
The presence of a delay does not mean there is no A.I. there. Not everything is as fast as Google search, for instance IBM Watson would think about a problem for a few seconds, which is fine because it only needs to be about as fast as a human.
Great point -- if a system has a good probability estimator it can send questionable answers for review to raise accuracy up to the commercially useful level.
AIs are already trained on huge corpuses of natural human communication, but they still aren't able to reliably analyze human language. Adding a few more gigabytes of chat logs isn't going to do much.
They weren't trained to read natural language, and then respond with some programmatic action filtered into natural language:
Bot reads "What's the temperature like near me?"
Person calls "$get user-local-temperature"
API responds "{temperature:{f:77},{c:99}}"
Human writes "It's 77 degrees outside!"
Training set now contains that relationship between that question, that API call, that response, and that natural language response (and probably the users location, age, gender, and so on, all captured in the meta-data about the response in the corpus).
Bot reads "What's it like outside?"
Person calls "$get user-local-weather"
API responds "{weather:{now:Sunny},{today:Cold}}"
Human writes "It's sunny now, but will be cold later today."
And so on. I think the goal here is training on standard API calls as the response, and taking their data return and converting it into grammatical sentences. It's a two step training process. Know which API to call, and know how to convert API response to natural language.
There's no serious corpus yet for that -- if this is real, it is important work.
Facebook's M is powered by Wit.ai (who they acquired a while back), which basically does what you're talking about. They translate natural language into machine readable data structures. Honestly, its interesting tech and being able to feed in facebook level data volume should prove very good for them.
thank you for being one of the few people who actually knows what's going on underneath. Wit.ai's software computes meaning via statistical modeling of semantic roles as they relate to "intent". It loosely resembles the linguistic theory of Frame Semantics (Fillmore), embodied today in the FrameNet project at ICSI in Berkeley.
The translation of unrestricted human language into a restricted, "logical" language which could be understood mechanically, is another long-standing problem - going back as far as Liebniz - which is unlikely to be solved by a few gigabytes of chat logs.
All true, but it doesn't need to solve that bigger problem to have succeeded at building a reliable M: it just needs a high success rate at converting requests into desired responses, not translating everything into some master language for arbitrary use.
Are you claiming that chat logs won't contribute at all? Or are you claiming that progress is irrelevant unless someone launches a 100% solution all at once?
I think the chat logs will help, but it won't just be about that data.
The data around how users interact will also be important:
Do they prefer a back and forth conversation, or do they want to say everything in one go.
Do they want to start a conversation, drop it, come back to it several hours later, or do they like completing it in one go.
How do they handle switching back and forth between different contexts, if certain requests take time, or do users not switch context.
What data are they happy to share, and what are they not.
What are the typical response times that a user considers acceptable 5 seconds, 1 minute, 5 minutes, 60 minutes? Does it vary depending on scenario.
Is there particular services or information that there is a trend towards, for example local search requests, research/information, particular types of purchase etc.
We've just spent 6 months going through a very similar process to this, which has helped drive the development of our Converse platform, which allows people to build semi or fully automated conversational messaging services, so this is fairly closely related.
(From our point of view, the NLP data we gathered was useful, but it wasn't the most important part)
Yeah I feel like some responses are pulled up, and then "M" gets assisted to pick the best answer, in the case an answer is not generated someone types one out. I would call it "Assisted AI" if anything, whereby the AI is monitored for accuracy. It makes sense to me, eventually in theory it would not need as much assistance, and even if it still would need assistance, it would still be one of the more reliable AI systems yet to some degree.
This is a new angle on the app-outsourcing-to-low-paid-contractors "technology": it's so dehumanising that you have to pretend to be a computer while you work!
It's also strikingly similar to the original "mechanical turk".
> "The opinion is split as to whether or not it’s a real AI, and there seems to be no way of proving its nature on way or the other."
Clearly, the author didn't even do the most basic fact checking. Since, Facebook clearly told everyone that M was going to be AI that was assisted by humans.
Yes, the article is explicit: "For more complicated tasks, such as making a driving test appointment at the DMV, the humans will do most of the heavy lifting. They’ll actually place a call to the DMV."
I think that also depends on who is actually completing/fulfilling the task, in the DMV example its AI adding context to a task that can't be automated easily.
The AI can still do the requirements/information gathering and leaving the job for a person to do. For common issues (support triage, customer service issues) even this still has significant value.
Equally from a training point of view, it may simply be the person correcting/confirming the AI is right, and leaving it to get on with the process, rather than fallover to the human completely.
I suppose the interesting point is more that it's being marketed as not really operated by humans as a positive when in many instance in our modern world the reverse is true. And also it's an interesting application of the Turing Test too. :)
Facebook's strategy here is to build an AI brand before they have the actual technology, which could make a lot of sense. At the same time the interactions between M's team and its users will provide meaningful data to train the AI on.
I've never heard vaporware characterized as something that "could make a lot of sense"; how could it?
Plus I'm doubtful whether the data would be very meaningful. A bunch of people adversarially trying to figure out whether the AI is real is not representative or generally useful data.
Microsoft Windows was vaporware for years. They famously did a "demo" that was just a manipulation of graphics.
But Bill Gates correctly grasped that the future of the business rested on it and set about building the brand.
How does your N=1 anecdote show that vaporware makes a lot of sense? I'd say they got away with it, but not that it was a crucial factor in their success.
Because that N became the dominant computing company for a generation?
Apple was way ahead of Microsoft with windowing technology. Gates even offered to help Apple port it to the x86 architecture. He was content with being the dominant application developer, not the OS.
Jobs blew him off so he seized the opportunity.
Around the same time IBM was working on OS/2, which was a much better technology than Windows. Microsoft worked with IBM on OS/2, but then gaslighted them by "debuting" Windows 1.0
Edit: And don't forget how me VC pitches are about "what we're going to build" vs. "what we've already built."
Vaporware is not an optimal strategy, but in many cases it works.
Look I can see that it worked out well for them in the end. What I'm specifically disputing is whether the vaporware-thing was a causal factor contributing positively. If they actually had the software already and didn't need to fool anyone, things might've worked out even better. In that case things worked out despite the vaporware, not because of.
I think if you have the foresight to see a multi-billion dollar opportunity, and you're behind schedule, then you do everything in your power to grab that opportunity, including vaporware.
If they actually had the software already...
No one sets out to lie. Gates would've preferred to demo the real thing. But in the context of that market, "we'll release that in two years" is worse than vaporware if you want to actually own that market.
I think the data would be very useful considering a vast majority of users won't be using it in this way, also consider that people who query for things will learn the limitations of the system providing the information and adapt their queries accordingly. If you have a system capable of providing meaningful results to highly complex queries then you can start bridging the gap between how people interact with machines vs how they interact with humans.
I just had the perfect idea for a test, but then I went back to the recent discussion on Mimic[0] and double checked my favorite example[1]. Google has already updated their support, but there is a chance that Facebook M is still behind. Test them now before it is too late:
"When is the next Τаylοr Ѕwіft concert in my area?"
Bad typing is definitely not enough to measure if an AI is really a human. As a teenager, I wrote a chatbot for an online text based game. I have it knowledge of a QWERTY keyboard layout, and when "typing" it had a small random chance of pressing the key next to the key it wanted to press. It would also sometimes transpose characters. Sloppy typing can be simulated.
Might be an interesting test to do a statistical analysis of your subject's mistakes against a corpus of real human mistakes, since there are many common mistakes humans make, and a random AI might make inhuman mistakes, but this would of course not be conclusive.
Simulating bad typing is only necessary to fake a human. Here, having bad typing when faking an AI is stranger.
That said, AIs trained through the chat transcripts of a large number of conversations may produce mistakes. I remember reading a paper that gave good results that way, with the side-effect that it produces typing mistakes as a result. I cannot find that paper again, unfortunately.
Clearly there are some humans behind M that are doing things that Facebook would rather entrust to humans (like making phone calls). However, the phone call only proves that this specific aspect of M is human.
In the end, though, I suppose it doesn't matter. I'm going to guess that the ultimate end-game on M is for Facebook to collect advertising/affiliate revenue from recommending things. For example, if someone asks for a Chinese restaurant, plumber, dentist, lawyer, etc. in their city, the one they suggest could be the one that paid Facebook for it. As long as these types of fees make it profitable for Facebook, it doesn't matter if the service needs to be powered by millions of humans. In fact, that would be great - it would mean millions of new jobs.
Larry Page famously told an early investor that Google wasn't yet sure how it would make money, but that search was the only situation in which people would tell a computer what they wanted, and that there had to be a way to make money from that. M is exactly the same - a way to get people to tell Facebook what they want, and it puts them in a great position to monetize it.
I wonder if you can ask M to use fewer exclamation points. From the conversations I saw in the article, it's a little too chirpy (or should I say "clippy") for my taste.
I had the same sentiment. Similarly - when I ask Siri what time it is on my new Apple TV, after midnight it always says Zzzz... as if it's judging me for staying up late.
I'm not a huge fan of AIs fake emoting all the time. Occasionally, it's amusing, but all the time it just rubs me the wrong way.
Thomas Edison claimed to have a long lasting light bulb before he actually did. He showed it to reporters one at a time in a booth. Between observers, he would change out the light for a fresh one. Source "How We Got to Now: Light" (on Netflix currently, at least in the US). Found the clip on PBS. Skip to 2:20 for the specific part: http://www.pbs.org/how-we-got-to-now/big-ideas/light/
He was also pretty bold on pricing, electing to set the initial price off what he predicted eventual costs of production would be. Initially, new products would be sold at a loss.
Perhaps not controversial in the perspective of modern venture-backed startups, but at the time it was a key reason GE won early market share on so many products.
He was a businessman, who hired people to invent things. He was certainly useful, but it's always been surprising to me how much credit he got for inventing things.
To a tiny universe of Reddit/XKCD readers perhaps.
In the rest of the world, for example at the offices of Con Edison, or in the city of Edison, New Jersey, he's still considered to be somewhat important.
There's a billion and one things one could mention about Tesla. Sadly, only the tiny universe of people who dislike Edison (and see him as a businessman, not an inventor) are aware of Tesla's contributions.
"Vaporware first implied intentional fraud when it was applied to the Ovation office suite in 1983; the suite's demonstration was well received by the press, but the product was later revealed to have never existed."
You are presuming that they will employ all of the people needed.
Why for example would they not turn this into a platform that easily added the AI benefits for external vendors and services, and just be the middleman collecting a cut in some way.
They are already starting to integrate vendors into Messenger directly after all.
> Can you tag for me all photos in my album that contain a kitten, but not a dog, with "kitten", and those that contain a kitten and a dog, with "pets<3"?
This approach is backwards. This is the kind of problem that is easy for a person, but not for an AI. So if M was an AI pretending to be a human, you could use this to determine that. But in this case, the suspicion is that M is a human pretending to be an AI - and they could simply decline to attempt the task, or pretend to be unable (or do a bad job deliberately), and you'd learn nothing from negative results.
I don't get the part about the reverse number lookup. Couldn't they be using a disposable phone number that is allocated to Facebook? That's what Handy, Airbnb, Uber, etc. do. Why would they have to block their caller id? And how does either method prove or disprove that M is human?
That doesn't prove that M's human, what prove's it's human is that a Human voice called. The fact it says Facebook is just evidence that it was indeed from M, and not just him getting his friend to call him and pretend.
> The fact it says Facebook is just evidence that it was indeed from M, and not just him getting his friend to call him and pretend.
It doesn't really prove anything, since caller ID is extremely easy to spoof (I used to call my mates from the emergency number for kicks when I was younger). Not that I have any doubt as to the credibility of the story.
This reminds me of the Focused people in Vernor Vinge's "A Deepness in the Sky", slaves that were integrated into the computer system to provide function that surpassed the computer's built in intelligence.
I've not read that. Personally i felt we're one step closer to the "cookies" in Charlie Brooker's "White Christmas" episode of Black Mirror.
From the article :
“Our test participant was impressed with how much M could do, but was sometimes disappointed at how long it took,” UserTesting’s report reads. “He concluded that it would be very useful if he could set it to perform a non-urgent tasks for him while he worked on other things.”
That made me shudder. One person tutting at the poor performance of "it". It seems plausible that robot-powered tasks would complete rapidly, and humans power the slower processes.
So the participant didn't know it was human-powered. If anything, that makes things worse.
To be clear, this is applying the classic observation that if you keep the first and last letter correct, humans are really good at unjumbling the center.
Wonder if you can choose a sentence such that humans see it one way (homophones, jumbles), perhaps via context clues, but the closest match (edit distance?) for the individual words gives a different sentence?
Certainly seems doable.
Along those lines would be something like: "if I have a coin and I, err,trun tit, which face is showing?" but it's not a good example. Here "err, trun tit" gets corrected to return but the end should find to "err, turn it" instead making the face showing be "the opposite".
Hopefully you get the idea, bet there are some really good phrases that would fit this scheme.
A better way to say what I was getting at is that fairly straightforward language statistics go a long way towards unjumbling letters. A spell checker could also include quite a few human perceptual quirks as scoring rules without crossing the line into what I would think of as training an AI.
Regardless of wether M is currently more human than AI, we could project that in the not-so-distant future (after it's trained), M will be mostly, 99% AI.
The technology itself will become more and more available and other companies will also use similar AI tech to work with customers.
The ultimate moment will be when the AIs start talking to each other in human language, each 'thinking' that the other is a human.
That will be the moment when the machines have decided something for you and while at first you'll think that you triggered that, at some point it will become unclear - is the human triggering the AI or is the AI triggering the human.
Pretty soon, everything we consume and everywhere we go will be controlled (and, a bit later, predestined and programmed for us) by the AI.
Based on punctuation analysis, word choice, tone, its not just any human but a mid 20s white female. Probably front ending google.
Real comedy would be going to mturk to try and find the task to communicate try to crack it recursively "M find me the mechanical turk task for this request".
What a silly conclusion. The fact that a human called his land line does not mean M (the thing in messenger) isn't an AI. At best, it proves that M has humans who work for M making phone calls.
I don't have any insight or opinion about the question of how human M is, but this article seems makes a bunch of assumptions that make the whole investigation somewhat silly.
So basically the suspicion is that M is a concierge MVP? From reading the chat excerpt I'd agree.
Edit: It would be interesting to devise a way in which you can make two Ms talk to each other (or have M talk to Siri etc.). Maybe "can you pretend to be a customer for my XYZ business"
As much as I appreciate the effort, I don't think proving M has humans behind it is any of help.
We write AIs. We try to make them act just like us. We teat them in everyway we can imagine and we expect them to act like a human would in response. Providing an algorithm for this is not always useful or maybw not even possible.
My theory on this is that Facebook is powering M with both people and some sort of AI software that not only analyzes and sometimes finds the best response, but it also analyzes the conversations people on both sides made.
Now this can be useful on several levels. Facebook can improve it's AI algorithm in less time, the AI can help people on their job in the meantime (by analyzing their work and commenting on it)
This guy is hilarious. If you're reading HN comments you know that AI isn't quite there yet right? We're easily 5-10 years away from anything you're looking for.
>The most noteworthy aspect of this reply is that “Google Maps” wasn’t capitalized, suggesting that maybe, just maybe, a human typed it out in a hurry.
Or they're smart enough to add random mistakes. When I started a project for setting up multiple ways to say the same form letter, I thought of adding a random-typo feature to make it look like humans were writing it. I'm sure these guys are at least as cheeky as me...
But they don't try to convince you M is a human. Indeed the opposite. So it would be rather stupid to add typos to an AI, when you want people to see it as an AI
I don't think there has to be a huge controversy here. It's perfectly plausible to build a system that contains a hybrid of human and machine intelligence, where the humans work on the more fuzzy questions that cannot be directly answered yet, and the interactions used to fill in the gap as the AI is improved for later.
My photos were blurred and I found in F12 developer tool that they failed to load. After a little fiddling with my network they successfully loaded and became clear.
It's not just you. It's basically unreadable on Firefox on Android (and what's worse... the page prevent arbitrary zooming, unless you "request the Desktop page").
It's depressing how a supposedly well-designed platform like Medium still falls short of providing an usable mobile interface.
