 Yeah. Light banter recording his own. Yeah, I was going to say that's a little annoying, but just to let everybody know we're recording now. We are here with my craft founder Joshua Montgomery. He's going to do his little light banter for a few minutes and then we'll while we wait for a few more people to join. And then we're going to hear about voice technology and why my craft is important. We do have a chat if anybody wants to dump a few questions for Joshua to answer. So hello, Joshua. Hello. So I've praised an Imam and a rabbit walk into a bar and the rabbit says, I'm pretty sure I'm a typo. Another one with the two horses are standing in a field and one horse looks at the other horse and says, I'm so hungry, I can eat a horse and the other horse says Is it the dad? At least I am used myself so anyway. I was going to say, now the sure father you have nothing but dad jokes. Yeah, there you go. So it's a beautiful, it's a beautiful Friday and it looks like we've had a bunch of folks join us. Mark says he missed the end of your last joke your favorite joke. Moo. It's a beautiful Friday. It is Friday. Thank goodness it's that time of week. I think it's Friday. It is Friday. We had a great meeting with most of the MyCraft team that ended earlier this week. Yeah that actually was really good getting everybody together in the same room and spending some time doing some higher-level architecture stuff on the on the software side and then spending some time working through the manufacturing on the manufacturing side so all that was positive. So so Chris it looks like we've got quite a few people here and did you want to get started? Please do. And can you remind me what the topic is for today? The topic for today is generally whatever you feel you want to discuss but I thought that a really great topic to cover today would be voice and why MyCraft is important. We've been hearing a lot recently about privacy issues and how MyCraft can help solve some of these problems that people are having with their devices spying on them. Okay well I think we should start with the renaming of the company or renaming the company effective tomorrow. We're going to call it MetaMeta which is more meta than the other meta. It's self-referential so that makes it fun and funny. I'm very thankful not to have a brand that's as damaged as the brand that Meta's trying to replace. I can't remember the name of the other company kind of slipped the tongue now that they've replaced it with the with the corporate name Meta it just it slips the mind to all of the behaviors and the other stuff that that company was involved in. So anyway you know as you know Facebook and other companies start building you know the metaverse right and they start building these online experiences that are truly immersive you know one of the things that they're going to be doing is learning a whole lot about the people who are using it and you know it didn't make a lot of sense at the time for Amazon to launch Prime Video because really they weren't in that in that business that they launched it and you know I kind of scratched my head as to like why they would get into into the production of movies and TVs and was chatting with a friend and you know we came to the conclusion that you know you can learn so much about people from their preferences and movies and entertainment that you know it looked to us like the the launch of Amazon Prime was a mechanism for our friends at Amazon to learn more about their customers and so you know launching all these these video services launching the metaverse yesterday you know in every case it seems to me to be a a ploy or a tool to allow these companies to access more information about about folks and sell their attention and then of course sell them products and services either from the company directly like our friends at Apple or you know from third parties like the folks at Amazon so so anyway so as the the integration between the real world and the virtual world continues to as things continue to get closer and closer and more and more meshed together you know it becomes increasingly important for there to be options for all these various services that do provide privacy and allow people to pay for what they get and not be forced them to pay in a currency that they they wouldn't normally pay with you know the for me at least that means you know when i when i pay for a video service i i just want to pay for the videos i don't want to watch a 30 second ad every every five or ten minutes you know for me that means on the the voice side of things you know paying for a voice assistant that represents me as the as the person using it and not the company that shipping it so that's what we've built at my craft and and that's what we're we're aiming to ship in a significant way later on in in 2022 so you know the the next question is this kind of what is the future of that product that we're shipping and and we've been having some pretty deep conversations and actually this week we we had several really good conversations about the upcoming roadmap and and you know for us it looks like it's a path towards you know a software focused company that does very very little in hardware it looks like it's a path towards you know integrating into other folks's devices as opposed to shipping devices ourselves and then it looks like it's a path towards you know empowering individuals and companies to deploy voice without being beholden to you know our servers and infrastructure being able to to put it entirely on the premises and so you know that's very much where we're headed I think there are there's some noise from some of the other providers in that that direction I saw that there's an integration going on with Walt Disney that includes a bunch of tech from Amazon you know but you know one of the things that that we're really privileged to have as a company and is a reputation for being private and and I think that the the competitors that we compete against really are going to have challenges about telling a privacy story and you know even if you rebrand it Meta at Facebook is going to have a lot of trouble convincing people that they're going to respect their privacy and the same for a lot of the other Silicon Valley companies and you know voice assistant tech is looking like it's going to be integrated everywhere right in in home offices in kitchens in garages certainly but also in corporate offices in retail environments in automotive in leisure and travel you know we recently did a project with NASA working in the aerospace sector and in all of these cases you know companies and individuals are looking for the ability to build a voice assistant that solves a problem for them that solves it in many cases in a way that keeps their their individual or corporate information private and you know they're looking to expand you know the coverage of the voice technology onto manufacturing floors onto cruise ships and and so on and so forth so you know the next question is is what does that mean for us as a company and and I think that it means that we have a very clear path towards success you know the the we've been working for a number of years on the core technology we were doing voice assistance before Amazon and Google made them cool you know we we shipped the mark one a couple years ago we're in the process of getting the mark two out the door you know supply chain supply chain willing and you know have had a lot of information or a lot of inquiries from you know all kinds of you know big manufacturers and manufacturing consortiums and government agencies and you know schools and others so it looks like that the market is really is really ours to address and you know and it looks like the you know the team is on the right track towards towards getting us there our plans for next year are to start shipping 5 000 units a month when the manufacturing lines starts and then start shipping significantly more after that as we work towards building a software solution and seek partners that are interested in integrating mycroft in their own smart speakers in their own automobiles and and aircraft and cruise ships and hotels and and other places so so yeah that's kind of where the future lies for us and in terms of the experience you know the that j that kind of artificial or janky back and forth experience that that we have with smart speakers today where you say the wake word wait for a response give it a single query it may or may not understand the context you know you have to go back and ask the same question over again that is clearly evolving towards a much more natural conversation both within our stack as well as the stacks of our competitors you know five or ten years from now I think that conversations with voice assistants are going to be much much more natural than they are today and 20 to 25 years from now I think it's going to be very difficult for people interacting with voice assistants across their domain of expertise so you know talking about things like music and pop culture and you know anything on wikipedia I think that folks will be able to have conversations with their voice assistants that that seem natural enough that that folks are going to have trouble determining whether or not it's a it's a voice assistant in some cases and that you know future experience opens a lot of doors for these technologies to be deployed everywhere really and you know for us it's it's really important that there be an option for for companies and individuals and you know communities that that might not speak a major language as their primary language for those communities to have access to voice technology that speaks their language protects their information provides them with privacy and so you know that's what we're building here at microbe so anyway that's a nice little primer on you know where we're going I'd love to talk to the community a little more and kind of answer some questions and address any concerns that folks have I know that we do a lot of broadcast in our messaging so we're on email and chat and you know we do email or we do blasts through the press you know I've been on a lot of podcasts lately but haven't had a lot of back and forth with the broader community so so yeah I'd love to open the floor to questions and I don't know if Chris is going to moderate or how I believe you can just unmute and ask a question it's up to you if you would like um Roger Austin just asked will mycroft be used for commerce will it accept payments and does mycroft have a payments partner so that might be um something that he's specifically interested in but um you have had some thoughts about payments in the past did you want to share those really quick yeah so um the the skills framework that we've built really is designed to allow both individuals and companies to deploy things easily um and to deploy new capabilities easily and and one of the obvious capabilities for the for the smart speaker stack especially given how much uh marketing uh amazon's put into it with Alexa is is buying things right whether it's goods or services or whatnot um we do accept payments today uh at least for using the back end of of the mycroft stack uh and we we use a I think an off-the-shelf payment partner for that uh but you know as we move forward I think that there are going to be a lot of additional opportunities for commerce within within the voice assistant space you know um I think one of the the cool things about the concept of user agency which is something that we really strongly endorse is the concept that the voice assistant that you talk to when you work with with mycroft is represents you as the customer and not us as a company right and so you know if we build a commerce a commerce skill that allows you to you know add things to a shopping list and then deploy those shopping lists to vendors to have have things delivered uh you know at the at the end of that experience our goal would be to have you know that shopping lists be parsed out to various different companies based on your best interest um and have those products show up to you at the the lowest cost um in the highest speed possible um not you know dividing that shopping list in a way that makes mycroft um you know that feeds mycroft's uh needs right so you know if paper if paper towels are less expensive at you know through walmart fulfillment than they are through target fulfillment and you know even if we have an agreement with target you know I'd like to see a skill best represent the customer and send that order across to walmart um you know that type of shopping experience I think doesn't exist because of the back end relationships that the existing platforms have and I think that there'll be a lot of advantages to it you know um always the lowest price right because the voice assistant represents you um we do want to accept I think payments and and ideally sit between the customer and the vendor um someday uh and and today we don't have a partner for that um but it's something that that we'd certainly be interested in looking at you know there's there's obviously a lot of work to be done around fraud prevention and you know verifying the veracity of orders and so on and so forth but it's something that I'd love to see as part of the mycroft stack um the generic mycroft stack and then of course if retailers out there are looking to deploy a smart speaker um you know into people's kitchens and into people's homes that allows them to order from that retailer so for example if the target wanted to ship a target branded smart speaker that allowed you to shop with target um you know they can absolutely use our technology to do that and you know one of the places where we see big opportunities is in helping some of these bigger retailers um to compete head to head with you know with amazon specifically um both in the u.s. market and abroad so hopefully that answers the question okay I hope so too um we had a couple more and I wanted to jump in with the next one mark asked can you speak more on mycroft sales and marketing efforts sure so uh our strategy today is attraction rather than promotion right and so you know one of the one of the things that we did last year as supply chains tightened in covid made things difficult um for for us to ship product um was we went from taking a full deposit on a mark two to taking a dollar deposit and and the reason for that was you know every time we took a full full payment for a device you know it created an additional liability for the company so you know we took money from a customer we owe that customer a product right so so it sits on the um it sits on the liability side of the balance sheet and we had so much demand that we just kept digging this giant hole of liability for for mark twos that needed to be delivered so we decided to to reduce the deposit to a dollar you know that way we we at least know that the the customer is is serious about buying the product because they opened their wallet and got out their card and gave us a dollar um but we're not creating a giant liability so you know from a sales and marketing perspective you know step one is is attraction so you know making sure that that my craft is part of the stories that that the broader media is telling about privacy both in Silicon Valley and in this space and specific and we've so far avoided mainstream press primarily because our user experience isn't at a level that we're really ready to have a mark to sit next to a alexa or a google assistant um and you know be evaluated head to head with companies that have spent you know in some cases thousands of times more money than we've spent building their technology stack but as our user experience continues to improve and and i would expect by you know late first quarter you know third quarter at the latest next year that the tech team will be across the line to providing a solid experience across the basic skills that we we plan to support out of the box um i think that there will be a lot more opportunities to to be part of that that broader media story you know the the second place i think that we have big opportunities to to market our product is across you know this entire ecosystem of the internet that is just abhors the the issues with the privacy have have uh the lack of privacy have created so you know advertising on duck.go you know advertising on new shows and on forums where you know folks who are are like-minded are you know looking at that crowd that's unplugging from facebook and giving them other alternatives and then finally you know one of the things that's happened that's been a real blessing you know all going all the way back to the the Super Bowl ad in 2016 where Alec Baldwin introduced the amazon amazon alexa is that big tech has created a lot of awareness and and demand for smart speakers and and so you know we get to ride that wave a little bit because you know people realize that they want to use a smart speaker and that this is something that they they want to have in their home or in their business or in their hotel chain or whatever because they see what what's happening with you know the capabilities that it brings to the table but then they examine it a little closer and they say oh wow the privacy implications here are significant you know i i'm not really excited about allowing big tech to spy in my kitchen or in my bedroom or in my you know my hotel chain and then they pick up the phone and or they pick up the internet search and they say you know private alternative to alexa or siri or google assistant and they land on our page and they come to our contact form and they send us information and we get a ton of those on an almost daily basis from companies that are super small in some cases up to you know cruise lines and hotel chains that have tens of thousands of rooms the other thing that that you know we bring to the table is mycrafts rapidly becoming the standard for privacy in the industry i was on the phone with a group called koala out of europe that is building voice technology for european industry you know they've got they've actually raised i think three or four times more money than we have and have this huge international team in the european union that's that's building things on mycraft to run in manufacturing plants alongside workers right so a voice assistant that lives in a in a headset in their case that allows workers to collaborate better with robots that type of activity you know whether it's in europe or whether you know we did something similar with nasa you know whether it's in educational settings like one of our customers chatterbox in all of those cases it allows us you know being part of the standards being part of the the the research that's ongoing about privacy you know creates big opportunities for us on the sales side of things because there are a bunch of big companies involved in that you know the ones that pop to my mind and koala whirlpools one of them that's that's doing a lot of work with them as well as a bunch of european companies whose names totally escape me right now so attraction rather than promotion i think is where we are today and then long term you know i see really big opportunities for us to be the white label for other brands that are deploying that they want to deploy voice so you know and we've had these conversations with many many many brands that you know they wanted to play a great example as we had a conversation with some of the folks over at nike early on when we weren't quite ready about the concept of taking various different sports celebrities capturing their voices and including a smart speaker with some of the the higher end sneakers that the nike ships so you know they they ship sneakers that are 350 400 500 dollars you know there's enough margin in there to ship a smart speaker with it and if that smart speaker comes with the the voice of a celebrity you know that that creates an opportunity for nike right and you know their customers use it for all the same things you normally use use smart speakers for but maybe they get a because their owners they get early opportunities at some of the the limited edition speakers that the the shoe company ships that type of application really you know it isn't possible today with with our friends in seattle or our friends in the valley now they may build something like that that's a real possibility but for many of these brands you know they they already have a really tortured relationship with silicon valley tech and they're looking to deploy those types of applications with companies that orbit them so you know the the way i kind of unpack that is you know when a company like world pool for example does a deal with google you know it's inevitably world pool that's in orbit around google google's the center of gravity because it's so much bigger and so much more influential in the global market when a company like world pool does a project with a company the size of mycroft you know we orbit them and we serve their needs and so i think that that's another opportunity on the sales and marketing side where you know companies that that are attracted to the technology that do you know even the most basic searches for privacy and user agency you know land with us and then you know our sales team works to close them as a as a corporate partner and and to make sure that they're able to ship our technology in their their own products in you know a way label smart speaker you know in their software you know wherever they choose to use it so that was a very long winded answer to sales and marketing that i i was just gonna say we're getting a lot of questions in here if you don't mind dropping i'll keep it shorter no that's okay the next question is what is available to use now on my desktop and before you get to that really quick i had one that i thought was important to kind of jump to the top um michael asks please define smart speaker so could you do both of those please okay so a smart speaker it's really interesting you ask that question because the uh you know we we came into the market before smart speaker was a thing and so when we originally started talking to outside investors we had to explain it to them like you know it's a speaker so it plays music and sound but it also listens so it's got a microphone in some cases it's got a screen as well to display information and when somebody chooses to interact with it you know either by waking it with the wake word or the context of their conversation the speaker responds with some some type of action and so you know in a broader sense you can think of a smart speaker as an augmented reality overlay over your the room that you're in so that was you know i was kind of joking about renaming the company meta meta but we could right except for facebook could probably sue us off the off the the edge of the planet um you know a smart speaker provides a virtual reality overlay over your room only instead of virtual reality the way that or augmented reality the way that people sometimes view it like it's going to be like the google glass thing where you know you look at bob and it shows his vital statistics next to his head right and and gives you access to an information layer it's actually an audio overlay it's a it's a device that sits in the corner and listens and when you want to access internet information you speak to it and it interprets that and comes back with a response and that response might be something like playing music or accessing news or showing you the time or setting a timer but of course those responses also you know integrate the internet with the real world right so you can turn on lights and lock doors and you know control vacuum cleaners and all kinds of other tasks that that really require you to access the internet and you do that seamlessly you do that without a screen in many cases and you do that with voice using natural language so so anyway that's how i would define a smart speaker in terms of how you use that on your desktop today it depends on what kind of desktop that you have on the linux desktop you can download and install the the microft application and unpack it and pair it and it'll it'll run on your your linux desktop on windows and mac i do believe there are docker containers for microft that work i will point out that we've been really careful as a company to stay away from anything that would make it easy for broad scale adoption right so the technique you only have one opportunity to make a first impression and you know the second that we build a windows installer or an ios installer or an os 10 installer you know it opens up you know millions tens of millions hundreds of millions of potential customers to us and you know if those customers download the technology stack and it sucks that's really our our one and only chance to make an impression with those customers so to date we've kept the microft stack more focused on developers made it not deliberately made it difficult but haven't invested resources to make it easy for for average everyday users to to use the technology and the reason for that is to preserve our reputation when we are ready to deploy on desktop on mobile on other folks's platforms will create packaging for those platforms and make it super easy to install but as very for today because the user experience continues to be developer pretty you know we've stayed away from making it particularly easy i was going to drop in i've got several more questions here scott asks do you see value in smart voice devices that combine a general purpose voice agent with a custom or application specific microft agent potentially along the lines of and there's roman numeral seven industries voice interoperability initiative i'm not familiar with that yeah there's there's a lot of initiatives going on to help make voice assistants work together you know and we're part of those and i think at least one of them has adopted our our stack is the standard through which they're going to be they're going to be standardizing some of the interfaces i think that there is a lot of a lot of value to making the smart speaker or making the end terminal dump or agent agnostic at least because it it creates the opportunity to create a ubiquitous overlay right so if if i have a friend who's just really into google assistant for whatever reason he's got one of those things in every single room he's even got one on his front porch so you can actually walk up to his house and like say hey play scary music on all the speakers at like three in the morning and it'll play it'll play him on every speaker in his house uh but yeah uh you know when i walk into his home you know i should be able to say you know hey microft you know play my favorite very playlist and and have that work for me you know you know using voice identification and and you know other other other tools to determine that i am who i am we're not quite there yet and i and i think that there's some really deep privacy concerns that need to be addressed before something like that is built but certainly i think interoperability is important you know i think that standards are important provided and this is a big provision provided that they're not set by one tech tech company to their own benefit right like everybody wants to set the standard to whatever they're doing and force everybody else to come along you know i think that the tech companies are really good at that and so i think that we should look very carefully at many standards that are adopted to make sure number one the company that's proposing them doesn't have a bunch of patents that make it impossible for other people to use the standard or that if they do have patents that by using the standard software they implicitly license them and then number two that the the standards are built in a way that makes them supportable by the broader community i really do concern myself and when it comes to standards of big tech that tech will set standards that force everybody onto an upgrade an update treadmill where developers have to constantly push updates all the time in order to keep up with the standard you know that that really puts potential competitors on a back foot and gives the the folks that are setting the standard a significant advantage so i think we want to pay pay close attention as those standards develop to make sure that tech doesn't play its its game so anyway chris you you had an i hopefully that answers the question chris you had anything you had another one yeah i've got a few more hold on let me get to it uh louise asked any news on the mark two you alluded a little bit about that but can you get us a really quick mark two update sure so we've been chipping dev kits for a little while now the microsoft team was actually here going through the non-technical team members were going through the exercise of building mark twos from scratch and so that we could pay closer attention to how that that construction process will work you know the design is done we selected a contract manufacturer last month we announced that it's it's aztec out of malaysia we're really excited to be working with them we're now trying to secure adequate parts and get our supply chain stood up to the point where we can start mass production and then importantly keep mass production going right now we're having the same supply chain issues that the rest of the world is having so you know things ridiculous things like little 70 cents 70 cent amplifier chips are unavailable and are making it difficult for us to stand up the line but i think that the supply chain disruptions will settle out in the next six months and so we're aiming for the third quarter of next year for shipping but that is dependent both on supply chain and then on funding so we're we're out raising money as well all right and next question josh asks are there any plans to include a specific piece of hardware to allow mimic 2 to function on iot low power devices locally like josh sounds like a developer yeah so mimic 2 is a speech synthesis engine that we trained using the voice of crissal the intern who is now i i kid you not an army ranger good for crissal uh the mimic 2 engine uses a machine learning approach to to rendering speech i know that they have been working on getting it working on lower power devices they've also been working on produs prop prosody cadence and tone a lot of that work has actually been going on within the mozilla community as well as on the speech recognition side in the koki community and as you we look at those technologies as they develop they they appear to be getting more resource intensive not less resource intensive over time and so you know making a better prosody engine that uses machine learning run on a smaller chipset is certainly on the roadmap we're not there yet someday soon i hope and then you know if you do if you are within our community i encourage you to ping uh ken uh ken ken smith uh who's one of our developers uh down in sarasota florida uh ken's been doing a lot of work on making microp work on low resource stacks just kind of in a spare time uh you may ping ken in the forums and ask him hey you know where is this and and if you have the ability to help you know how can i help uh because i i know that we'd love to see it and uh i know that ken's students have worked in that direction okay the next one is from uh simon he asks since you are focused on values like privacy rather than maximizing profit have you ever considered becoming a purpose corporation uh yeah big corporation yes is the answer um uh yes you don't want to elaborate just a while okay yeah please well no i i can i can say two things yes we've considered it um yes um i i think we even have drafts of the language ready to to incorporate into our into our corporate charter if we haven't already i would expect that to happen uh at some point in the future uh and i'm actually scratching my head wondering if maybe that actually got done because i know we had lengthy discussions about it uh the you know we've always regardless of what the actual corporate organization we've always been purpose driven and and i don't think that that's actually independent of profit right so you know if we abandon the principle of privacy and we abandon the principle of user agency then we're just a really really really underfunded competitor to alexa and siri and google assistant right um i think the the place where we have an advantage is that all of the silicon valley tech companies have damaged their brands so thoroughly around privacy that they can't credibly address the issue and so you know even if our friends at google had just the the the most sincere intention of providing a really truly private experience for the google assistant and maybe they do nobody would believe them right like a vast majority of the public just wouldn't believe it because of their past behavior um you know in our case i think that that creates a really great opportunity so our focus on privacy really is i mean it's definitely a moral stance but it's also a market stance you know as as we look at the market the research that we've seen says right around 20 of the smart speaker market or the potential smart speaker market is sitting on the sidelines you know because of the privacy issue right and that's a huge huge market i mean you're talking about billions and billions of dollars in sales that are sitting idle because you know people simply don't trust the big tech companies so as our manufacturing comes online you know i think that that's our our addressable market and that that means a total addressable market size of something like three billion dollars uh and growing every year and so i i think that the the privacy aspect is actually our competitive advantage and it is something that we should continue to to talk about and of course to implement because you know it's great great to talk about privacy but you have to actually provide it and so that's one of the things that we do and and we'll continue to do all right roger asks is it possible to partner with open phone os's like lineage or sailfish seems like that would be a community that would appreciate voice that is private yes and we'd like to do more of that we're limited in resources is why we have it but yes the all of the open phone os's are our real potential distribution mechanisms for us and i think that they're going forward there are big opportunities in being on mobile devices that aren't running uh vanilla android or for ios uh scott asks do you see value in adding edge on device voice technology to your software stack for example user authentication using voice or commands that would that work even if you lose internet connectivity uh yeah and actually at some point we'd like to move the entire experience to the to the edge i mean the the cloud experience that we provide now kind of comes back to that question about mimic 2 earlier which is you know it just takes so much resources to to process a full language speech right that it's just really unaffordable to put that on on a device and you know and in addition to being kind of a technical challenge from a cooling perspective and so on and so forth uh so yeah the the as silicon continues to get cheaper and more powerful as we learn more about how to pack these models down and make them run and with fewer resources moving the entire experience to the edge i think is the most private experience possible and and so yeah i'd love to be able to do it honestly today and provide people with the smart speaker that sits there and the only time it goes out to the interweb is when it needs to obtain data to answer a query other than that everything is local you know i think that that's the best the best way to provide privacy for individual users in their homes but then of course you know that's also a great solution for using voice assistance in boats right or in cars or in aircraft or in spacecraft or you know in remote locations or on job sites you know places that might not have a ubiquitous internet connectivity you know having a voice assistant that runs on device would be a really really powerful tool so that that's definitely in the future it's just a matter of of both waiting for morse law to help us solve the problem as well as once again resources i had a feeling you were going to bring out morse law on that one next question is will you establish certification requirements for your smart speaker using mycroft to assure the quality of the user experience for example false rejection ratio maximum false rejection ratio false alarm ratio and response accuracy uh yeah i mean as hopefully we're crossing our fingers that one of the manufacturers out there will eventually come around and and bang on our door and say hey we want to make a smart speaker that includes the mycroft stack you know when that happens if it carries our brand there's going to it's going to require the certain performance metrics and so in that case you know we'll will require a certification you know we are an open stack though and so if people want to take our technology and use it wherever they want to use it and they don't use our brand you know they're free to do with that as they choose but if it has you know mycroft's logo on it and the goal is to provide a a consistent experience you know regardless of the touch point whether it's a mobile device or a smart speaker or a desktop or a car or a boat or whatever all right next question is what is your business model with oems wanting to integrate mycroft nre royalty support yes you want to elaborate a little bit all of those right like um you know the what we've found is we've worked with um work with other companies that are looking at mycroft is they do need to support in order to make it work and work well within their within their technology so so that's a place where um where we can be helpful uh you know there's generally going to be a non-returning charge to help get them stood up and help get the you know all the the specs built and help them to to figure out how to integrate this into their product and then in an ongoing basis you know a licensing fee for you know back end software and other things that are licensed in a way that requires a payment so we are an open source company but our licenses are designed for designed in a way that if you're going to make commercial use of the product you need to have a conversation with us well that is all the questions i have in the chat if any other attendees want to add a question or even um ask yourself now would be a great time and joshua is there anything from we it sounds like we've got a lot of really great developer focused questions today is there anything else you wanted to add no i'm i'm i'm really upbeat on the future of what we're doing you know the the we've got there's two types of hardware companies i wouldn't call us a hardware company but we have to make an initial version of the tech in order to have a reference device for people to to play with and evaluate and you know in some cases white labeling use you know there's two types there's the type that have a warehouse that's full of things that people don't want and the type that has an empty warehouse because they've sold all the product you know from day one we've been the second type right you know everything that we've produced has gone out the door just as fast as we can produce it so i'm really excited about taking that and instead of producing you know hundreds or thousands of units a month you know eventually producing you know hundreds of thousands or millions of units a month either you know with us as a software component of somebody else's speaker or you know us producing reference devices natively so you know i view it as a a big opportunity you know i view it as a largely untapped market you know the the privacy market for smart speakers i i challenge you to point at something that that you know rivals us in terms of its ability to to provide privacy and also you know provide a meaningful a good user experience you know and i'm really happy with you know the team that we've built we've got you know great folks on the on the technical side we've got 60 000 plus developers inside the community that are doing work in various ways on the stack we've got great leadership from michael and you know we've got great support from folks like chris and sarah and others that that work really hard to to make all the gears in the machine move so i'm extremely optimistic about our future you know big market great product you know big barrier for other folks to to enter and at least into our segment of the market and and you know first mover advantage you know all of those are moving in our direction all right i have a couple more questions that have come in michael asks uh so please please supply an overview of how microth works on my desktop and the implications to privacy sure so microth on the desktop um you know it runs an agent that that opens up the microphone uh the wakewards are processed locally so we don't actually have visibility of them in the cloud i when it detects the wake word it takes a short sound sample and sends it up for transcription uh that transcription is aggregated across all the devices so the the the machine learning uh system that is actually doing the speech recognition doesn't have visibility of where those queries are coming from uh we feed back a uh a text transcription of what was said and then the natural language understanding actually takes place on the device once the device is figured out what you're trying to do it either takes an action like you know open a web browser window or you know turn on a light or you know whatever you have a program to do uh or if it's response to speech it'll send the speech um the text up to a speech synthesis engine and we synthesize the response send it back as audio and it places out the speaker so all of that takes place very very quickly um you know the there there are privacy implications there in that you know it is calling back and forth to internet um we have done the best that we can to make sure that those are our those interactions are encrypted uh we log nothing beyond the we log nothing and keep nothing beyond the length of time we need to service the queries so there's no history associated with that uh and you know with us in terms of privacy we are an opt-in organization if you want to share data so you know we do ask our the folks using microsoft if they're willing to share data so that we can improve the machine learning models and improve the user experience for everybody but there's no requirement to and and by default we don't take any of that data so it's only people who affirmatively say hey i want to share data that we we keep that data and then finally there's transparency you know you can look at every line of code in our software and examine exactly what we're doing with the data and how we're handling it and you know if you do spot a security vulnerability or a privacy problem it'll exist uh only as long as i'm unaware of it and uh you know openness and transparency are really critical to privacy because you know as we've seen with with the Silicon Valley giants many of whom are under indictment in the european union by all of the european union nations by the federal government here in the united states as well as all 50 states and it takes quite a lot of work to be indicted by basically the entire government across all of its layers you know these black box software packages that they build oftentimes don't do what they say they do and they don't treat private data the way they say they do and so openness and transparency and the ability to look inside it is really critical from a privacy perspective um roger asks any plans to put mycroft into automobiles yes just as soon as we have an automaker that wants to do it we did take an investment from jaguar land rover and we did do an integration with the f-type sports car but that was very very early in our development and importantly that was with their us team and you know if you've worked with jaguar in the past you realize that the the center of gravity within that company is in the uk and so despite building really good relationships here in the us to actually get built into the car requires having great relationships in the united kingdom and for whatever reason personnel or communications that that never happened but it does you know it does highlight a great application for the technology within automobiles jaguar did make an investment and you know as our user experience continues to improve we plan to circle back there as well as with the providers of technology to the automaker so folks like bosh who you know build a lot of the parts that go into cars the other place that we have a lot of interest and if anybody has a connection over there we'd love to we'd love to talk to them you know it came to my attention last week that our friends at tesla have a music streaming service um which i guess they've licensed from a third party but is included with their vehicles and i think it would be a lot of fun to build a tesla branded smart speaker that goes you know maybe they give it away free with the test drive or you know shows up with your with your car and allows you to interact with the vehicle from your kitchen you know provides all of the other smart speaker aspects and then uses teslas music streaming for for music you know i know that they have some voice technology inside of that vehicle today but i i suspect that given the amount of compute on that vehicle on those vehicles and the full-time network connection that there are some real opportunities for building voice assistant tech into them and you know the other guy who's launching spaceships has a voice assistant you know i think that there's there's a lot of value in bringing one into the tesla ecosystem so if anybody has a good connection at tesla i'd love to have a chat all right i think we might have more questions than we can answer in the last 10 minutes but let's speed round we got eight minutes speed round so michael asks so no log can be obtained we don't log anything beyond the length of time that we need it to provide the query so the query comes through the computer we transcribe it or take action on your behalf if you have not opted in to share your data it is deleted and that's that okay nick asked is server part of mycroft open source and available to the public it is it's called selene um the selene architecture is all available it's fairly well documented um it does take several vms working together to do that to provide that service but it is all open source that is all available and are getting a repo all right roger asks could mycroft be an encrypted phone slash messenger system if so how would it compare to other privacy phones yes you could probably integrate signal into mycroft without too much trouble and as with many many other things with mycroft you may get into the community and and look and say you know is there a signal integration for mycroft and find that somebody's already built it i'm just unaware of it that happens all the time somebody comes up with an idea and it's like oh yeah like some community member built that a year ago so i'd have a good i'd go and have a look at it i don't think that will ever be a phone operating system but i think that we can be a nice layer on top of a privacy phone that gives people the ability to use modern voice tech in a private way and i think that signal would be a real good back end for that both for the voice and the messaging josh asks what are the plans for mycroft on android either full stack or ui recording on device and processing remotely probably the second and actually not probably there are people who have it working in the second bay including that koala team in europe that are using a 10 and a half inch samsung tablet that interestingly on their slide was 10 and a half inches but 400 grams which totally confused me because i thought they were european centimeters um yeah there's been a lot of work doing doing getting mycroft working on android and as far as i know there are several implementations out there uh eventually i'd like to obviously move the whole stack on to the under the phone and i think that morse law is going to get us from here to there sooner rather than later especially as some of these tensor processing units and other things become available um so yeah i mean android is is an obvious place where we'd like to be once again we do want to be careful not to deploy an android app to the google play store that doesn't work well because it it has the potential to do more harm than good but once we cross the line to a user experience that we feel comfortable you know making it available to the general public you know an android app is definitely in our future michael asks how do you protect from manufacturers slash china backdoors in chip design i don't think we're using any chinese chips i think we're using taiwanese and us chips primarily but yeah you know supply chain attacks supply chain attacks are something that's difficult to defend against you know using reputable suppliers with the reputable manufacturer is really the best we could do at our size as we get bigger i i think that's something that we need to pay attention to because i agree there are there are definitely potential downsides to using chips that were made in may munch on them and simon asks despite the improvements mimic still sounds quite robotic can you talk about its future development and how to make it sound more natural sure so the original mimic uses the allen pope voice which is a concatenative approach and is very robotic mimic 2 uses a machine learning model that was trained by chisell the intern who i mentioned earlier and sounds a lot better but even mimic 2 was built several years ago now and the state of the art for speech synthesis using machine learning has improved drastically in the last three years and so you know i think it really comes down to taking the chisell data you know feeding it to newer algorithms and then if necessary obtaining additional data for that for that speaker i think that the the other piece that i'm looking forward to seeing is using transfer learning to train these models so using a model that sounds good from a prosody cadence and tone perspective but then layering a voice on top of it using transfer learning i think both of those are approaches that can significantly improve the quality of the voice in terms of the the quality of the voice and the machine learning feedback loops and some of the other stuff that we're trying to do it's primarily resourced resource constraints that keep us from doing the work we we just have only have so many developers and even with the big open source community you know it's very difficult to solve some of these problems you know open source developers are awesome and they work hard but a lot of times they deliver a 90 95% solution and the remaining 5% is on the full-time staff and you know we have very limited full-time staff that is all the questions i have and so i just wanted to see if i couldn't get you to share a little bit more about where the community can learn about mycroft and learn about our reg cf sure so you know we're out raising money one of the things that we believe in as a company is that everybody who contributes to the company one way you know whether they're a developer or somebody who buys one of our speakers or they just mention us to a friend should have an opportunity to invest that we shouldn't be beholden to one or two silicon valley venture capital firms we would rather have a broad community of small investors so we've been raising money through crowdfunding since day one when we did our first kickstarter today we're raising an equity crowdfunding round about five million dollars just a shade under that through start engine chris can drop the link to the start engine campaign in the in the chat here you know the we appreciate anybody who's interested in in kind of taking this roller coaster ride with us into the future we see a big success over the coming years and you know we see big opportunities in the space for folks like us who are focused on privacy and user agency and some of the other things that silicon valley has has failed to materialize thank you joshua and i think with that we will go ahead and end it we had a a nice hour long conversation we will host another one of these we're going to have one with our designer derrick and he's going to talk about why the design of the mark mark two is the way it is and one with michael louis our ceo and he's going to talk about roadmap so stay tuned thank you so much thanks folks have a great day