I’m a big fan of Halloween. I had over 100 kids come to my door last night displaying their creativity in both the themes and execution of their costumes, and I was happy to reward them with a little candy. But this season is known not only for treats, but also for tricks. Unfortunately this year has brought us both. Let me tell you about it…
As you may recall, we were expecting to receive the first article circuit boards from the board house last week. Instead they called us and were asking for clarification about placement of some of the components. Their questions weren’t really making sense, and after a few rounds back and forth it dawned on us — they’d built an old revision of the circuit board. The problem finally got tracked down to human error. A technician had simply grabbed the wrong files and they produced the wrong ones. So we’ve been reminded why you do first articles.
The board house has owned-up to the mistake and are correcting it. We aren’t certain exactly how long this trick is going to take to recover from, but I’m guessing it’ll be a couple weeks. This isn’t exactly a disaster like I spoke of in my September blog post, but it certainly is an unexpected detour on the road to shipping.
Despite this setback, we are pushing forward with our plan to open source the hardware design. I’m happy to announce the Mycroft Mark 1 now is on our GitHub under the hardware-mycroft-mark-1 repository! Everything you need to build our first Mycroft reference device will be found there.
Open Source Software is pretty well understood, but Open Source Hardware is still a fairly new concept. The licenses to support it are still being honed. We are using the CERN license, which is often called the “GPL of Open Source Hardware”. It was created by CERN in Switzerland, the folks who brought us the Large Hadron Collider. However you likely won”t be able to build a super-collider — even if it is open sourced — unless you have a 27 kilometer wide backyard; and even if you did the electric bill for running a 6.5 teraelectronvolt experiment is prohibitive. Fortunately building a Mycroft Mark 1 unit is a lot more practical, and nearly as cool.
We will also be applying to become a member of the Open Source Hardware Certification Program along with companies like Adafruit and Sparkfun Electronics. We encourage everyone to produce (even for commercial purposes) or remix our design, just keep sharing any improvements with the world.
We’ve also been pushing forward on the software platform. On the forums we have published the Mycroft Roadmap for 2016-2017. This document highlights the many areas we see as important technical targets to strengthen today’s Mycroft platform. Portions will be tackled by us here at Mycroft Inc, but we also want and need to work together with the community to tackle many of the tasks.
Additionally we’ve published the Mycroft Skill Lifecycle on the forums and have begun to fill the mycroft-skills repository on GitHub. This provides a great source of working skills and examples of how to hook Mycroft into home media and IoT devices, and more.
Please pop over to the forums, read, and join in the conversation!
I received a request for an email interview from Jasper, a student in The Netherlands. He had what I thought were some really great questions and I thought other people would find them interesting. Perhaps not quite as riveting as Interview with a Vampire, but hopefully my answers are almost as interesting as his questions were!
What’s your ultimate goal with Mycroft?
Create “AI for Everyone”!
- We want to make voice technology pervasive, with the same Skills (or “voice apps”) available to the user wherever they want them. Might be a phone, a desktop, a watch, or car or a coffee maker that they are talking to, but I still want to be able to ask what is on my calendar.
- We want to allow people who can’t use technology today because of discomfort with traditional interfaces — like the elderly — to be able to just talk to technology.
- Similarly, we want to empower those physically limited, such as someone paralyzed or chair-bound
What part of the software do you think is holding back further progress? Do you think it’s possible to create a ‘perfect’ voice assistant?
Do you think being open source is an advantage or a disadvantage for Mycroft? Does Mycroft gather a lot of interest from developers and hobbyists not affiliated with the company?
How trivial do you think it is for Mycroft to have deep third party integration? How do you plan to achieve such integration, and do you work together with other companies to make sure your software works?
What do you think about other comparable technologies like Amazon Echo, Google Voice Search and Apple’s Siri?
How do you think AI will influence our lives in the future? Do you imagine everyone will have a dedicated piece of hardware in their homes to use voice assistants or AI?
Steve has been building cutting edge yet still highly usable technology for over 25 years, previously leading teams at Autodesk and the Rhythm Engineering. He now leads the development team at Mycroft as a partner and the CTO.