Echo, the First Useful Home Computer Intelligence?

We’re familiar with features like Siri or Microsoft’s Cortana which grope at a familiar concept from science fiction, yet leave us doing silly things like standing in public yowling at our phones. Amazon took a new approach to the idea of an artificial steward by cutting the AI free from our peripherals and making it an independent unit that acts in the household like any other appliance. Instead of steering your starship however, it can integrate with your devices via bluetooth to aide in tasks like writing shopping lists, or simply help you remember how many quarts are in a liter. Whatever you ask for, Echo will oblige.

Screen Shot 2014-11-06 at 2.57.14 PMThe device is little more than the internet and a speaker stuffed into a minimal black cylinder the size of a vase, oh- and six far-field microphones aimed in each direction which listen to every word you say… always. As you’d expect, Echo only processes what you say after you call it to attention by speaking its given name. If you happen to be too far away for the directional microphones to hear, you can alternatively seek assistance from the Echo app on another device. Not bad for the freakishly low price Amazons asking, which is $100 for Prime subscribers. Even if you’re salivating over the idea of this chatting obelisk, or intrigued enough to buy one just to check it out (and pop its little seams), they’re only available to purchase through invite at the moment… the likes of which are said to go out in a few weeks.

The notion of the internet at large acting as an invisible ever-present swiss-army-knife of knowledge for the home is admittedly pretty sweet. It pulls on our wishful heartstrings for futuristic technology. The success of Echo as a first of its kind however relies on how seamlessly (and quickly) the artificial intelligence within it performs. If it can hold up, or prove to hold up in further iterations, it’s exciting to think what larger systems the technology could be integrated with in the near future… We might have our command center consciousness sooner than we thought.

With that said, inviting a little WiFi probe into your intimate living space to listen in on everything you do will take some getting over… your thoughts?

129 thoughts on “Echo, the First Useful Home Computer Intelligence?

    1. It can hear anything in your house? Better hope your intimate partner’s name isn’t Alexa or that you don’t unintentionally say their name. Amazon sells gags, right?

      *shakes head*

    2. YES! FInally! This is a feature I called for in a 2010 white paper and I’m very happy to see someone is finally trying to make it happen.

      The next step would be to integrate it as a wearable, with camera. Then you could say things like “remind me tomorrow I have an appointment on Friday,” or more complicated things like while walking through a store you are looking at an aisle full of stuff – asking “which is best for my needs” will cause it to help you decide by engaging you in conversation. And staying there long enough will cause it to offer to get a person for you to talk to about it. Then you could pick up a document in a foreign language and say, “Read this to me in English,” or you could share contact or to-do information with your peers or a girl you met at the store by simply asking, “what’s your name, blahblahblah, could i call you some time?” and she says “sure; call my cell” and the contact negotiation is automatic.

        1. It’s cute in TV shows, but star trek plays in a rosy future where people have gotten rid of some bad habits. In our present reality if you put this in your house you should in fact be removed from said house and be put in an insane asylum with no delay.

      1. >The next step would be to integrate it as a wearable, with camera

        we could call this device “a phone”

        > Then you could say things like “remind me tomorrow I have an appointment on Friday,”

        we could call “her” Siri, or Cortana, or just demand google does it NOW

        >or more complicated things like while walking through a store you are looking at an aisle full of stuff – asking “which is best for my needs”

        then it checks its list of sponsored products advising you to buy one from a company that paid the most

        >Then you could pick up a document in a foreign language and say, “Read this to me in English,”

        it reads “Dear Aunt, let’s set so double the killer delete select all”

        >girl you met at the store by simply asking, “what’s your name, blahblahblah, could i call you some time?” and she says “sure; call my

        pimp…

    3. My first thought as well. I 100% guarantee the unit isn’t locally based, I’ll guarantee its “cloud” based. There is no way that small of a unit, for that cheap would have what it takes to do the computations, and storage. Speech recognition is a massive amount of algorithms. Check out big blue and Watson and right there is quite literally the most advanced technology of speech recognition, and quite literally as well the EXACT definition of what amount of computation is required. Here comes the thought police

  1. It reminds me of an old firewire iSight camera turned on it’s end and anodized black.

    While this effectively brings my childhood fantasy of having the ‘Computer’ from Star Trek:Next Gen. into reality, I really really don’t like the privacy implications that go with it.

    Make a standalone internet-less version that relies on a local database of a few terabytes of useful information and I’d be more inclined to own one.

    1. But then how would it work at that price point without the “marketing research” of keywording and indexing your daily chatter and conversations with housemates/friends/family/noise from the TV?

      1. Alexa, stop tracking the shows I am watching and selling that information to others or using that information to sell me more stuff. “I’m sorry Dave, I can’t do that.”

    1. This hack is how something that promises lots of (blank) to idiotic consumers for low amounts of money, and they’ll buy anything, irregardless of the terms and conditions, freedoms, or whatever else they lose, so long as blank makes them a) feel empowered b) lets them be lazy and not learn something c) lets them narcissistic(ie domination) d) lets them be greedy(self-righteous or selfish). Its called hacking the human population. And in my personal experience it works the majority of the time rather than not, not being usually 25% or less.

  2. I absolutely love this idea! My main concern, though, is this: Where does the voice processing take place? I have a Windows Phone 8 and if I don’t have network access, Cortana can’t help me. Now, obviously Cortana will look up the search term (or whatever) from the internet so I have to connected to do a search. But, if I turn on airplane mode and tell Cortana “set an alarm for 15 minutes” I get back a message saying “I’m sorry, the internet and I are not talking right now.” Seemingly offline tasks require an online connection, which I can only assume is for voice processing.

    So, if Echo/Alexa’s voice processing takes place “in the cloud” and it waits to hear the activation work, it’s not just possible that it’s a privacy concern, it *is* a privacy concern.

    1. I’m positive it throws recognition into the cloud (like Siri, Cortana, and GNow). I’m hoping that it’s hackable enough that we can replace the pieces with local alternatives (like Sphinx speech recognition) or a custom “automation processor”, but it’s probably unlikely.

  3. you actually featured a hack a few months ago about turning a rasberry Pi into exactly this same thing for $150 dollars less with open source hardware and software. So, if going back 9 months into the past is bringing the future, than this does bring the future. https://hackaday.com/2014/04/09/create-your-own-j-a-r-v-i-s-using-jasper/

    I’m not usually a troll and I do care about new innovations, but this is being covered everywhere and it doesn’t really represent an innovation for the hacker community — UNLESS — amazon makes the technology opensource.

    1. Exactly. I don’t know why Jasper isn’t even mentioned in the article. Hacker-wise this is the much more interesting alternative. And the topic of ultra home automation is still a rather geeky one – I think for most of the potential buyers Jasper is a thing to look into.

  4. Not a hack, but it’d be a potentially interesting thing to hack… I have no desire to broadcast my every conversation to the cloud, but as an inoffensive hardware box that has decent microphones and speaker and a wifi and/or bluetooth connection, I can see it could be repurposed quite nicely…

    1. They are going to use an on board DSP to listen for an activation word. It should only pass up the speech to the cloud after you give that keyword.
      As an inspiration for a hack… A raspberry PI with camera and mic in each room. Open CV to see who is speaking. Microphone plus a good set of speakers. Sphinx to find the keyword to start listening and then pass along the speech to you home server running say an i5 or AMD A series cpu to do better voice recognition using sphinx. Then a natural language parser. Throw in some home automation and you are in good shape.

      1. “It should only pass up the speech to the cloud after you give that keyword.”

        “should” is the word in question here when it could easily be sending every we say back for them to data mine.

        1. yep but frankly that would pretty useless. Too much data and processing with no real gain. Also they are a company so they would be open to all sorts of legal action if they used this as a wiretap.
          I trust the desire for profits and the fear of billion dollar court cases in this case.

          1. There’s nothing illegal about it. If you buy a product that advertises it is listening 24/7 and that’s what it does, you can’t sue them. And the amount of processing isn’t as intense as you imagine. Simply logging every word and ranking them by frequency would be invaluable to advertisers. Imagine if you could know how many times your brand name was spoken in the average house.

      1. Why would the government buy information from Amazon. They already have enough information about you. And seriously, so what if people hear you. I know it changes a bit when you are in the privacy of your home, which is supposed to be considered private, but if they do what they truly say and only listen for the keyword at first, whats the big deal?

        1. Why would the government buy the data from Amazon when they can just claim it under PRISM for free? Of course you should be worried about that. Things said innocently or in jest could be taken in an entirely different light later and a false case could be constructed against you with your own words that the government was able to obtain without a warrant. Our government has massive data retention capabilities and computers are very good at indexing and keywording your speech.There is really no difference between having an Amazon Echo in your house and a Commisar.

          1. A commissar has eyes to catch you writing notes to each other that Alexa can’t hear!

            Can Alexa be re-named? Does it have a record of all searches? Does it give answers that depend on your previous questions? Can you zone it (answers from say Ireland rather than USA)?

            No, I don’t want an early version; V2.2 perhaps, depending on reviews.

  5. While we’re at it, why don’t we just replace peoples’ brains with a computer as well? It would quite easy – all it would have to say is “What?” and “I don’t understand” and “Where’s the tea?”

    Hopefully you’ll get the HHGTTG reference there….

  6. Great I guess I won’t be needing my brain any more, maybe I can sell it on Amazon. That will make it much easier to stop worrying and learn to love the surveillance state. “Alexa, who should I vote for in the next election?” “You should vote for dear leader in the next election, he loves you and you love him.” Good enough for me. What is it going to cost me for this magical tube?

  7. The “intelligence” in the headline is right. Ponder for a moment what would have to be changed to make Echo better at spying on conversation in your home or office. You’ll draw a blank. It’s on all the time. It has speech processing to pick up what’s spoken across the room. There’s no hardware lock out as with Siri.

    We have to assume that it only comes fully on with it hears that magic word. But keep in mind that it is, at the very least, listening for that word all the time. And while I doubt Amazon has a spy mode built-in, It is almost certain that several of the world’s three-letter spy agencies will soon be looking for a way in.

    1. If I was as paranoid as you, I’d probably be living off grid in a faraday cage. It looks to me like I might also implement the mute button, use the power switch, regulate internet access to the device if I were concerned about the nsa spying on my boring life.

        1. Because understanding how things work makes you see how easy it is to hack things. Hence Hackaday. The layman sees a router and it either works or not. HAD users see how easy it is to login and steal information going across the network. Hence the paranoia and conspiracy theorists. Most of the suff isn’t that far fetched.

        2. This is not conspiracy, the single largest purchaser of zero day vulnerabilities in the world is the NSA. Is everyone really so blind as to think that the government bypassing our constitutional protections by purchasing mined data from corporations is not a problem. I would expect better from people using this site. It’s really sad how willing people are to hand over everything they say or do to the very people who will do all they can to throw you in prison when you step out of line and/or piss off the wrong person.
          If you can go 1 week without breaking a federal law, you are in a very tiny minority of the population. This is what they want, so that when/if you ever decide to stand up and be counted on an issue that means something, they can lock you up and throw away the key.
          Just look at Bernie S. from 2600 or Aaron Schwartz if you want an example.

      1. umm..its not paranoid these days we KNOW the CIA does this.
        They recorded the phone calls of a whole country,they installed their own hardware in routers,they installed their own software/hardware in facebook,google etc.
        Just googles around, this was all big news stories you must have missed.
        How people can think that Amazon is immune is baffling.

        As for your life being “interesting enough” – we know they look at nude photos from peoples emails, and stalk ex girlfriends. You hand thousands of people limitless spying power at the push of a button OF COURSE they will use it for random human vices.
        You ever cut them off in traffic? They might look up your license plate.

        1. Do you have any examples of this happening that concern you? I’m sure there are loads of people in positions of power that could destroy me and they don’t need to wire tap my house to do it. Why would they?

          Why are some of you so concerned about my personal decision to sell my privacy to a 3rd party?

          Furthermore if the agency can already spy on all of my traffic how does adding yet another device change anything?

        1. Not the processor, but cells in general;
          http://abcnews.go.com/blogs/headlines/2006/12/can_you_hear_me/

          Dunno about 2 decades, but its certainly been there awhile, and thanks to Snowden we know this stuff is now routine to the point of being piss easy.

          Also worth noting the new facebook app specifically asks permission to listen to your mic and read your text. They have excuses though – reading your text saves you from typing in a password reminder, for example.

  8. This thing… It reminds me of Eddie the computer in the Hitch Hikers Guide to the Galaxy. Preferred the 1981 or the audioplays of 1978.

    The more intelligent the stuff around us, the more knowledge is lost. For example the father in this demo of Amazone rather asks its new ‘computer’ than to think for himself.

      1. oh yes surgeons bring books into the operating room with them so they don’t need to keep the information in their heads

        using crutches all the time for everything means that you start to rely on them for everything

    1. “The more intelligent the stuff around us, the more knowledge is lost.”

      This crops up time and time again, yet I bet your life is massively more complex then it would have a few hundred years back.
      Some knowledge will always be lost as skills and required jobs change, but theres zero evidence that the human capacity for information has gone down and nor that out total knowledge is less.

      We are infovours – we seek out information. Its why the internet is so addictive. You can question how useful a lot of what we absorb is (a lot has always been “gossip”), but at the same time we are exposed to so much more information – instant news from around the world, the addiction of browsing over wikipedia, social network feeds with updates from friends and the off forward petition again whatever.

      Also, surgeons have computers, x-rays nearby and in some cases google glass displays these days. – they get up to date information on the patient and details of exactly where to cut (sometimes with projections).
      Its not remotely that they arnt skilled or know what they are doing – its so they get vastly more accurate information then its possible to remember.

      1. Agreed, I wouldn’t have the access to the same knowledge without the internet. I use computers, internet etc. And I don’t have to remember everything. Its all about knowing where to search, in the old times no-one would read every book in the library and remembering every part.

        I see around me how my non-technical persons adapt and get used to smartphones and 24/7 internet access. Maybe addicted is a better description… I don’t have the solutions but this thing won’t get a place in my house.

  9. Yet another not finished hack laying around having the same goal as Echo:
    It’s a Hasbro macaw parrot toy-robobt, I’d like to improve (massively) it’s speech recognition and also being able to control it’s (limited) move.

    Why? Frustration about voice recognition ! You don’t expect a parrot to understand you perfectly, and it’s also put AI in another perspective: you’re talking to a “parrot”, what kind of intelligence do you except from it? Animal Intelligence ;-)
    Wasn’t there someone who did research in this field?

    Probably, I did wait about something like the quad Arm 1.6GHz Radxa Rock Lite board… or smaller.
    Raspi and voice recognition ? Anyone here?

  10. “With that said, inviting a little WiFi probe into your intimate living space to listen in on everything you do will take some getting over… your thoughts?”

    Sounds like a mobile phone. People keep those in their homes, right?

  11. Q: If you do get an invite and you don’t have prime can you sign up for the 30 day trial and still get the prime discount or not

    Regardless still skeptical its as good as the promo shows, no way is this worth 200 full msrp is my guess. Normal price should be closer to 100 with 75 or so for promotion

  12. Great! Another sponsored hardware that people will dump on ebay soon.
    And when Amazon starts to sell them off, we’ll already have a hacked bootloader and an alternative firmware. Maybe it’ll be suited as a DLNA loudspeaker.

    1. agreed this screams overpriced for basically a bluetooth speaker 200 bucks?? seriously this will be dumped really fast after the novelty as warn off and people realize they paid that much for a overpriced toy; even at 100 promotion its too much

  13. First thing that I did was search for an API but I didn’t find one. Imagine turning this into a home control device.
    “[name] Can you please water the lawn today?”
    “[name] Please turn off all of the lights in the den.”
    “[name] Please switch into security mode.” ( where if it detects voices and doesn’t hear a deactivation phrase it turns on all light and notifies 911.)
    [name] Please remind me to take [a medication] every six hours for the next month. Keep reminding me until I verify that I have taken it.

    What if you could network two of them?
    [name] Did my mom (elderly, in a different city) take her medications today?
    [name] Can you conference with [friend’s name]? (Basically speakerphone but with no phone service, only internet. )

    1. Cool. Though it does show that the industry is lagging behind. This thing (featured in the article) is pretty much just a siri in a standalone version. It does what any android phone can do with the right app.

  14. “freakishly low price Amazons asking, which is $100 for Prime subscribers”

    A cheap Chinese android phone or tablet can do the same for less surely?

    That said, I guess decent microphones ramp up the price here.

  15. I’m really sick of people passing off speech recognition coupled to a search engine as “artificial intelligence”. Oh, does it have a couple dozen hand-coded special features, like calendar scheduling? Big deal. Compuers still don’t /understand/ anything, and the technology for that really hasn’t advanced significantly since the 80’s. Watson? You mean the shiny new hardware housing 30 year old software concepts?

    Cortana, Siri, and Echo are interesting novelties but not artificial intelligence by a long shot. Skeptical? Then ask your favorite one to do something simple, like “find a pot big enough to cook mom’s potato soup recipe she emailed me for everyone who has rsvp’d to thanksgiving dinner.” These toys are little more than an awkward and slow interface to a handfull of common websites. It just goes to show how little natural intelligence the average consumer has these days.

  16. Since this thread has turned out to be a reasonably sensible discussion, please allow me to point out the trivial;

    “An OBELISK, from Greek: obeliskos,[1] diminutive of obelos, “spit, nail, pointed pillar”) is a tall, four-sided, narrow tapering monument which ends in a pyramid-like shape at the top.”

    “Pillar” maybe?

    With literal machines listening to us we should be very careful what we say.

  17. At last !!!!!
    I’ve always dreamed of little Big Brother on my desk and watching/listening me all the time.
    1984 comes real. At least things can only get better and better from now on.

  18. I for one, welcome our new overlords…
    Speaking of literals, a true test of this sort of interface would no doubt be double or triple entendres. Anything less is Dragon Dictate in a box.

  19. Nice way to plug your paid advertisement, wheres the disclaimer? HaD promoting a device that is a 100% invasion of privacy. Not only dd you get bought out but you sold out as well. Shame on you.

  20. The Echo is out for a month, I’ve had mine for two weeks, and it’s astonishing. I have no relationship with Jeff Bezos or Amazon; I’m simply a Prime Member. The free NPR news, the free 100 or so internet radio stations, free music and access to my Amazon Music Account ( which can import iTunes for a fee) is worth the price alone. It works so quickly, the response time to questions seeming to have no lag at all. The shopping and to do lists work accurately. The dictionary works flawlessly (for instance, telling me that Aussie is correct: Echo is a column or pillar, not an obelisk; the Washington Monument is an obelisk).

    Although I do worry about the recent baby monitor hacks occurring with this device, as it is always listening (only active when one says Alexa, but still, it’s listening), I do have an Android phone which is listening for “okay, google now”. So, for those that worry about Big Brother, although I share your concern, there is little way to avoid it unless one places their cellular phone in a soundproof box every time they’re not using it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s