MycroftAI – Update from the CEO Part 1 | Hacker News Reader

MycroftAI – Update from the CEO Part 1

community.mycroft.ai

31 pointsrickoooooo11 comments

Disappointing, but not unforeseeable. As a backer, investor and owner of two units, I'm sad this ended up here. Hopefully the project assets will be useful.

If you figure out how to make use of the hardware, please let us know. Getting a unit I couldn't even install another skill on made it functionally useless to me. I don't need a robot to tell me what time it is, I need someone to add carrots to the fucking shopping list or turn lutron caseta lights on when I'm rummaging around with my hands full.

Even if it's tearing it apart and repurposing the guts for a frisbee, I'm looking for ideas.

I am not surprised by this announcement. I backed their kickstarter in 2018, thinking that I would receive it within a year.

With that mindset, I got their software running on my desktop and started to contribute. I quickly realized that they did not want the community to participate in development of their core platform, as evidenced by the fact that it took them weeks to pull requests that fixed very basic bugs. In a very short time, their demonstrated apathy destroyed my interest in contributing any further.

I received unit almost five years later, after a complete redesign. I have moved onto other projects and the whole project has left a bad taste in my mouth. As a result, my unit is presently up for auction on eBay, but I might be stuck with it.

If that turns out to be the case, I will look into donating it to a developer working on another open source voice assistant project. Nominations for that potential recipient would be welcome.

And while I’m ranting on this topic…

Speaking as a veteran embedded engineer, I think their biggest mistake was outsourcing the building of their hardware. At the very least, they probably would have saved money had they have hired an experienced in-house team.

Regardless, I think they should have focused solely on the software and enabled others to put the stack on custom hardware. I can’t imagine how much money they pissed away on the hardware, while the basic stack still needed gross improvement. They spread themselves too thin by being jacks-of-all-trades, so they were masters of none.

MycroftAI will likely be ceasing development by the end of the month. I've used Mycroft a bit myself and even developed a few skills for my assistant at home. While it wasn't as good as the Google Home or Alexa in my experience, I really appreciated an open source option that had user privacy in mind. I was rooting for them to eventually build something very functional. Hopefully someone will take up the mantle and there will be other options. It seems OpenVoiceOS might be a good option.

This explains why home assistant (nabu casa) was able to hire the rhasspy (self hosted local only voice assistant) developer after he got hired by mycroft in 2020.

Considering the lack of adoption in voice assistants even _inside_ the "walled gardens" of the device manufacturers, this seems a bit...inevitable.

For some reason I thought mycroft was dead like a year or two before 'Rona.

It's basically on life support for 2-3 years now.

I think I see the problem here: https://mycroft.ai/product/mark-ii/

$500 buys an M1 Mac mini btw.

Definitely more than most would spend on a voice assistant at the moment. Also, the design probably leaves something to be desired for many.

Reply on news.ycombinator.com