Echo, Meet Mycroft

The Amazon Echo is an attempt to usher in a new product category. A box that listens to you and obeys your wishes. Sort of like Siri or Google Now for your house. Kickstarter creator [Joshua Montgomery] likes the idea, but he wants to do it all Open Source with a Raspberry Pi and an Arduino.

The Kickstarter (which reached its funding goal earlier this month) claims the device will use natural language to access media, control IoT devices, and will be open both for hardware and software hacking. The Kickstarter page says that Mycroft has partnerships with Lucid and Canonical (the people behind Ubuntu). In addition, they have added stretch goals to add computer vision and Linux desktop control to Mycroft.

With or without Mycroft, people are going to hack things like this together. If you dream of being able to start your teapot with the command “Computer. Tea. Earl Grey. Hot.” then Mycroft might be a pretty good leg up on getting started. We’ve also seen Echo integration with Roku and even Nest. We imagine an open platform would spawn a lot of interesting hacks. You can find out more about Mycroft’s plans in the video below.

50 thoughts on “Echo, Meet Mycroft

    1. I agree that the way she says the commands makes it seem a very arduous and tedious thing, like talking to those automated phone services.
      And the name does indeed remind you of MS, even if you also think of sherlock’s brother

      “Possessing deductive powers exceeding even those of his younger brother, Mycroft is nevertheless incapable of performing detective work similar to that of Sherlock as he is unwilling to put in the physical effort necessary to bring cases to their conclusions.”

      Odd choice of name perhaps?

      1. No need to assume or guess it is stated on the page why the name in the FAQ Section:
        Why name it Mycroft?
        Mycroft is named in honor of Mike, the supercomputer in Robert A. Heinlein’s classic novel “The Moon is a Harsh Mistress”. Heinlein’s Mycroft was a High-Optional, Logical, Multi-Evaluating Supervisor, Mark IV, Mod. L” – a HOLMES FOUR. Mycroft’s friend Manuel named him “Mycroft” after Sherlock’s elder brother Mycroft Holmes. This was later shortened to Mike.

        In Heinlein’s book Mike became self aware when the Lunar Authority hooked so much hardware to it (decision-action boxes, additional memories, associated neural nets) that the system became self aware. This took place gradually over time.

        Heinlein’s fictional explanation in mirrors our own approach to achieving a strong AI.

        For those of you who might not know; there are two types of AI. A weak AI is a system that is obviously a computer. It can perform certain tasks and solve certain problems, but cannot initiate action on its own or hold a meaningful conversation with an educated adult. A strong AI on the other hand is capable of communicating with educated adults in a way that is indistinguishable from a human being.

        So how does one distinguish a strong AI from a weak AI? Alan Turing – the father of the universal computer – proposed a test known as “the Turing test” where humans are asked to judge a computer based on a casual conversation. If humans are unable to determine whether a program is a human or a computer – the program is considered a strong AI. The recent movie “Ex Machina” was an excellent dramatization of this test.

        By this definition Mycroft is a weak AI. It can interpret natural language, initiate search and maintain the context of a query, but anyone interacting with it for more than a minute or two will know it is a computer.

        Our long term vision at Mycroft AI is expand Mycroft’s skills and abilities to the point where interacting with Mycroft becomes indistinguishable from interacting with a human being. When that time comes Mycroft will become a strong AI – hopefully the world’s first. Obviously we won’t achieve this vision for years, possibly decades.

        We believe that the best approach to achieving strong AI is to add diverse skills to Mycroft one at a time. If properly integrated and coordinated these skills will gradually create an entity that is greater than the sum of its parts.

        Since this approach to achieving strong AI would take a proprietary software company decades of effort and hundreds of millions of dollars, we have made Mycroft open source. It is our hope that developers, hackers and makers will develop new skills and abilities for Mycroft and contribute them back to the greater Mycroft community. Gradually moving the system from weak to strong.

    2. Mycroft is the name of Sherlock Holmes’ older brother.

      Though Sherlock initially tells Watson that Mycroft audits books for some government departments, he later reveals that Mycroft’s true role is more substantial. While Conan Doyle’s stories leave unclear what Mycroft Holmes’ exact position is in the British government, Sherlock Holmes says that “Occasionally he is the British government […] the most indispensable man in the country.”

      He apparently serves as a sort of human computer:

      “He has the tidiest and most orderly brain, with the greatest capacity for storing facts, of any man living. The same great powers which I have turned to the detection of crime he has used for this particular business. The conclusions of every department are passed to him, and he is the central exchange, the clearinghouse, which makes out the balance. All other men are specialists, but his specialism is omniscience. We will suppose that a minister needs information as to a point which involves the Navy, India, Canada and the bimetallic question; he could get his separate advices from various departments upon each, but only Mycroft can focus them all, and say offhand how each factor would affect the other. They began by using him as a short-cut, a convenience; now he has made himself an essential. In that great brain of his everything is pigeon-holed and can be handed out in an instant.”

      — ”The Bruce-Partington Plans”

    1. Mycroft is “Always Listening”. Does this mean everything I say is being streamed to the Internet?
      Mycroft is always listening. When you call him, he answers … Otherwise he ignores you so your conversations are confidential. Mycroft respects your privacy.

      Mycroft only begins streaming when you call his name, all other audio is ignored. The process of analyzing the audio is done locally and not retained, until Mycroft hears his name.

      1. Then we would be living in In Australia, there was a software program based on the commodore Amiga 500 personal computer called White box which was used for phreaking, it used CCITT#5- (R2) tones to manipulate the phone systems in Australia in a similar way to blue boxing white box screenshot. In other countries white box can be referred to a Portable DTMFtone-dial keypad with speaker which was used to access an answering machine to hear your messages when you were away from home and also could be used on PBX phone systems that required tone dialling & used to generate tones if the telephone is rotary-dial or its keypad is locked.

  1. I completely agree this should be done without the big ‘friends of the NSA’, but even if he isn’t corrupted and he manages to deliver a clean product, he’ll only be bought out by google or amazon or some such and it’s back to square one. Except he’ll have a nicer bank account . Good for him, not much good for the rest.

      1. So you are living under a rock, and it’s such a big rock you find you can’t move it, and you had the sense to install internet all those years back when you crawled under it but you created a firewall that was too big to get any news.
        Good for you I guess?

    1. Same thing that happens with most university projects full of piss and vinegar out of the gate, just something to show effort and experience to employers, then real world says hello, then abandon ship and move on.

      Ideally its still there and you can by all means contribute all the code you want to it on github.

      1. In the comment section of Mycroft it says they are game to share with each other:

        Creator Joshua Montgomery on August 26
        We’re not using any of the Jasper project, but we would love to form a formal relationship with them to make it easy for us to share code and help each other out. If any members of that community are reading this, let us know! :)
        Otherwise, we will reach out to them in the near future and say “Hi!”

        1. You have a rather negative opinion of the jasperproject “full of piss and vinegar” … “real world says hello, then abandon ship”.

          Why Mycroft is any different? Is there any strong business plan? I understand that there are much hype right now, but what next year and after? Do you plan new generations of products?
          Google voice recognition are pretty accurate, even on non-english language! Any other solution is mostly a joke eg. in polish language or croatian. Will Mycroft try to address non-english languages? Is there a business plan in this area?

          I wish you the best, would be excellent if it could compete even remotely with google voice recognition…

          1. Its not a negative opinion towards the project, it was more towards uni projects in general. piss and vinegar just means “aggressive energy” like yay new project so many hopes and dreams, were going to conquer the world, etc.

            I actually hope they get some more energy for the project in the future.

            I just really want open source Linux speech recognition worked on as much as possible and feel sad when they peter out.

            I wish that androids local device speech recognition would get backported to Linux as it works not so bad as I rather enjoy stuff more that is not “cloud” based.

            What languages will Mycroft speak?
            The initial version will be English followed very closely by Spanish, German and French.

            Since it is an open source project, the number of languages is only limited by the willingness of contributors to translate the voice prompts into their own language. They expect to eventually support 10 languages including Arabic, Mandarin, and Hindi.

            as for competing with online stt that is not the point, they says they are open to using all of the stt and having multiple input methods.

            Who are you using for speech to text (STT) processing?
            We want to be transparent with our supporters, but at the same time we need to keep some details of our project private until negotiations are complete and contracts are signed. Remember that Kickstarter is not a store. It is a place to build support and access capital to complete amazing projects. Our platform is still more than 10 months away from release and we have a lot of details to iron out before we ship.

            We are currently evaluating several STT application interfaces (APIs). Our software is designed to use multiple APIs simultaneously. Partially this is to improve performance, but it is also to prevent getting locked into a single technology or vendor. When we’ve selected and executed agreements with our upstream STT providers we will communicate our selection to end users.

            We will also remain open to adding STT vendors in the future or bringing this portion of our technology in-house.

            To preserve end user privacy we are looking at several mechanisms to randomize STT query destinations, mask IP addresses and conceal other personally identifiable information.

            Mycroft is open source so users who don’t like our STT or AI selection can always deploy their own STT or AI back end.

    1. I contacted Mycroft about getting the sources, here is the reply:

      Right now our aim is to release the repos under GPLv3 early 2016. This is to ensure they meet a certain level of quality before we put them out in the world. Too many projects have released unfinished or unstable code too early and have been relegated to the “not good” category because of it.

      We are looking at bringing community members into the project to test the code early, so if you are active in the community there is that possibility. Also, the more money we raise via Kickstarter and Indiegogo, the earlier we might be able to release.

        1. Oooh, development costs money, waaaah. This is why it will never be “the year of the Linux desktop”. So many people involved in OS are anti-gov, anti-corp, anti-payingforanything, self-entitled Galluses.

  2. Yeah, i went on to their kickstarter. I always hate reading the project leads’ profiles. All of them are marketers. one guy is called the product designer, but he looks to be an industrial designer. The cfo is a system admin for a kansas library. So the tally is 1 guy doing software, one guy designing the box, one accountant and 3 marketers. Its pretty obvious where their priorities lie. No thanks.

  3. It’s sad that they raised so much money for this obvious failure-in-the-making. Alexa Voices Services is quickly opening up, and it’s going to blow this toy (along with Apple’s HomeKit) out of the water. This sort of BS gives a bad name to Kickstarter projects.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.