The Divorcing Apple Project: Siri Replacement

Well, the divorcing Apple project is almost complete. And it feels great to be rid of that abusive tech “partner”.

Any gaps in Linux, I was able to fill by getting pass-thru to work on an OSX VM using VirtualBox.

The thing I’m missing now on the primary Linux workstation is the voice recognition and Siri style computer commands.

Anyone found a viable replacement? Something that offers accurate voice recognition? And able to assign macros to voice commands?

(Preferably open source & auditable.)


Have you looked at Mycroft?

1 Like

Not yet. Your experience with it? Is it accurate enough to be viable?

“We promise to never sell your data or give you advertisements on our technology.”

…Sure… where have we heard that before.

I’ll trust the architecture and not the promise.
Is it possible to run offline? Or does it have to process the data on 3rd party servers?


I read about it and I’ll get the v2 whenever that happens. My own experinlence is limited, I tried it briefly back in the day on my laptop. You have to download skills for it (for example, there’s a skill to access Wikipedia so you can ask it questions) to gain different functionality. If you want it to do something that it currently can’t then you can write your own skill. It can run as a self-contained service on your computer. The voice recognition seemed accurate enough, if not Alexa accurate (probably better now though). Its not hard to set up and play with, I think it’d be worth your time to take a look and see if it’s suitable for what you want.

1 Like

INGSOC, a subsidiary of Oceana, has a product called Big Brother. Not only will they not sell your information, they will keep it to themselves!


You’re so 1984


Can someone please give examples when voice recognition assistants are actually helpful on a desktop? I know they have been around for quite a while but I could never think of many ways of using them for productivity rather than for just playing.

Dictation is perhaps the only one I can think of. Even this I don’t find useful as texts I want typed rarely get generated by my head faster than I could type them, especially when you often want to go back and review/rephrase what you have typed, insert certain punctuation an so on.

So, what’s the deal then? Like “open web browser, google blablabla”? What have I been missing?


Number 1 example: people with disabilities.

There are other times where it could be convenient to speak to your computer instead of walking over to it and typing.

For myself I see voice as an input method for smaller systems that can then reach out to larger more powerful systems and report back, so a raspi that is running a digital picture frame with a voice input instead of having to ssh or connect a kb/mouse or other equally niche solutions.

1 Like

In 1994, OS/2 Warp v3 had Voice Type Dictation (too primitive to be truly useful) and Voice Command, which worked well in Warp and would be very useful on the L5.

What Voice Command did was to make shadows voice activatable. Shadows were sort of like a Windows shortcut or Linux symbolic link, but were really a subclass object of a file. Imagine recording a sound clip, then being able to tie it to a file or link. The, when Mycroft (or whatever) hears it through the mic, it activates a corresponding script or cli program with full arguments. :boom:

1 Like

Here’s one for user1 when government agents are about to break in:

Auto Destruct Sequence Armed


They’re incredibly helpful for annoying the people around you. :rofl:

1 Like

A little foresight goes along way.

1 Like

(and @Greendrake)

From what I’ve seen … that’s just about everybody. :slight_smile: Seriously, it can be difficult to train non-technical people to use computers, even when the computer has a GUI. They might legitimately ask: I can already speak my natural language (whichever one(s)), why can’t I control the computer that way?

There are two interdependent capabilities that a computer needs in order for that to work.

  1. Ability to decipher sound into words.
  2. Ability to convert words into actions.

… and enough horsepower to do both with acceptable accuracy.

For myself, there are times when “hands-free” would be useful, but not essential e.g. when I am carrying too many other things.

So I would like to see more progress in the Linux world in this area.

“Open the pod bay doors, Hal.” – 1968
Are we there yet? :wink:

“Abusive tech partner”

It’s hard to take anyone seriously when they talk like this.

Did they beat you and threaten you if you left? No. You knew what you were getting into when you bought it and it seems you got out without any scars. And in fact you admit to willing using their OS now in a VM… I guess you still love them a little bit.

Stop the hyperbole.

“Open source and auditable”

Have you audited or checked the audits for all the software you own? And what are you auditing them for? At what level do you audit the important tech you use?

Do you need a hug?

1 Like

No, just working a working desktop and mobile technology stack that Linux and most (not all) continually fail to provide to me even when I invest in purpose built hardware and software from companies in the ecosystem.

I use lots of open source software on MacOS and Windows, and buy/donate to companies that do it right.

But seriously… abusive tech partner? sigh

I could launch into a privacy advocacy spiel, but I assume you know what you’re getting into.

1 Like

Of course I know what I’m getting into. There is a balance between privacy and functionality. Going with maximum privacy would make my life infinitely more difficult.

Its true. I experiment with how far I can go towards privacy and still function, but there are always unacceptable comprises. @Emily is more hardcore than we are, it seems.

We are. Though this is somewhat tangential as far as “for the desktop” in that commands to open the pod bay doors (garage) are generally more useful from mobile devices or from other smaller compute devices.