Slashdot Mirror

you are aware of mycroft https://mycroft.ai/

Damnit, that's actually not at all as billed, and in many ways just as bad.

It can't run 100% locally. It requires you to make an account with their servers and send your data to them.

Screw that.

Check out Mycroft for an open source voice assistant. I've got it controlling my house with its OpenHAB skill and it does the timers and things just like Alexa.

people using a device in an unsanctioned way then complaining that the door was closed on it. That's the risk you run playing with open doors you're not supposed to see.

No, that's the risk you run playing with a device that you don't control.

A better way: MyCroft + devices designed to talk to it.

Otherwise, live by someone else's cloud, die by someone else's cloud. When you give up control, the entire problem is: you gave up control.

Stop giving people money to own your ass, and they'll (mostly, except where the government forces them on you) stop owning you.

Do any other phones/cars/speakers have this option right now?

My Garmin nüvi 3597LMTHD GPS has done this since I bought it in 2013. It's not connected to anything, not wifi, not cellular, and not bluetooth. Where I live (the boonies) the traffic features and even the map updates are pretty pointless, so there's never been a need to connect it to anything. Yet it understands me just fine. And unlike Alexa and others, it allowed me to rename it — it only responds to "yo, bitch", which is just how I like it. :)

Is the unit's understanding of language in general up to par with todays systems? No. But does it work for what it needs to understand? Yes. Very well.

For the home, when and if MyCroft gains a local speech understanding capability, that's the way I'm going. Everything I want to do is local, and the unit can be customized to run just about anything you put together (of course, commercial products aren't that easy to figure out, but that can be done in many cases as well.) Everything that depends on the "cloud" has failures, comm losses, and security concerns. Local is definitely the way to go.

Otherwise, everything you say ends up sent to Google, Amazon, Apple or whoever. And whoever they partner with / roll over for / get hacked by.

I trust Apple a little bit more as they've been pretty clear about being privacy focused, but that door is open for them to do "whatever" with your data, and it is best to keep that in mind. If they go local, that'd be nice. But inasmuch as it's a closed system, whereas MyCroft is an open system... yup, still going MyCroft if they can pull this off.

I'd trust voice 'assistants' in about this order:

- Amazon

- Apple

- Google

- Just about anyone else

Brutal! Mycroft loses on trusthworthiness to a bunch of competitors that you definitely know for sure aren't even slightly trustworthy. And Mycroft works for you!

This means you trust people that you know view you as prey, more than you trust yourself! I bet lots of people are like that. No wonder we have a criminal president. Let me guess: did you vote for him?

What are /. opinions of Mycroft.ai? For those who've never heard of it, it is a commercial effort to create and maintain open-source software to compete with Alexa & co..

With less than $100 you can create your own personal voice assistant. Here are 4 free replacements for Amazon and Google's assistants:
Mycroft - https://mycroft.ai/
Kalliope - https://kalliope-project.githu...
Jasper - https://jasperproject.github.i...
Adrian - http://www.theadrianproject.co...

Mycroft and Hass.io may interest you.

It's been done but needs work I'm sure. A raspberry pi and Mycroft is just a recipe /.ers love. https://mycroft.ai/

I have no plans on get a smart speaker or make one. I'm not that worried about privacy. I just don't see a real benefit from having one. I like my terminal when using a computer. I like my wireless keyboard/mouse combo for controlling the computer attached to my TV. It's a bit more raw and I prefer it that way.

I am considering backing the Mycroft 2. I am not about to pipe all my audio to Google, Apple, Amazon, et al., but this seems like a fun toy. I passed on the v1 because it used Google's STT, but this one apparently has 8 different STT options, one of which is Mozilla-developed and can run on local hardware.

According to Fast Company, their business model is framed around selling voice services to major companies who are similarly wary about sending client data to Big Tech firms. (The for-example is Land Rover-Jaguar.) This seems reasonable, and it provides incentive for Mycroft (which is open source; in part? in full? I can't quite tell) to continue to play honest or risk the cash from the privacy-conscious corporate partners that they hoped to attract.

I'm not totally sold, yet. I'd be interested in /. views one way or the other, or anecdotes from anyone who has a v1.

Amazon Echo, Siri, Google, etc. is Mycroft. I'm not sure if it's quite mature yet, but they are making leaps and bounds: https://mycroft.ai/

Use Mycroft with a local speech-to-text server.

https://mycroft.ai/

Local STT with KaldiSTT https://github.com/MycroftAI/m...

I've been reading into this lately -- I do want an open source solution to this. I think the basic premise is quite useful.

Take a look at mycroft.ai. I haven't dug into it too deeply, and I can say that the speech-to-text runs through public cloud today due to a lack of current open alternatives - but the architecture is modular, open, and comes with (alpha/beta) hardware options if you don't want to go the route of mobile OS or raspberry pi.

Can't vouch for them, but I would probably be dumping a lot more time into them if I had it.

Things like these are the reason why I'm not donating money to Mozilla.

If - and I don't yet know if this is the case, they don't actually seem to say - this represents a stand-alone, does-not-go-to-the-LAN-or-WAN speech-to-text system... with an error rate of 6.5% on English speech as claimed... then it's way more important than Yet Another Web Browser.

This is precisely the kind of thing projects like Mycroft need to become not just another way to send your activity out on the net, which inherently decreases both reliability and security.

If indeed this is what this is, then the door opens for all manner of sophisticated home advances we can actually trust and depend on.

They claim around 1:1 [decode rate : normal speech rate] with a reasonably modern CPU/GPU. That needs considerable improvement. Reference quote from here:

On a MacBook Pro, using the GPU, the model can do inference at a real-time factor of around 0.3x, and around 1.4x on the CPU alone. (A real-time factor of 1x means you can transcribe 1 second of audio in 1 second.)

That's a lot of computing power to hand off, particularly in a laptop. Using just the CPU, you'll be pegging it the whole time you're talking, and then some. For a decent desktop, it's at least doable, but it's still a very heavy compute load.

Though... saying "MacBook Pro" doesn't really tell us enough... I have a MacBook Pro that is a dual-core Intel machine... it's not what you'd call quick. There are a lot of different hardware configs that could be described by "MacBook Pro."

Seems like a pretty big deal to have to dedicate a server to the STT task (but then again, if I could get my STT tasks out from under the cloud... I'd probably do it. I have a spare 3 GHz 8-core hanging around, so...) but I think for general use, they have to do better. This isn't going to fly well on a Raspberry pi, for instance, it'll just get way behind.

Still. IMHO, this may be important. Very.

Its simple https://mycroft.ai/
Keep your digital assistant open source and you never have to worry

Mycroft.ai lets you host your own speech-to-text server and do everything locally if you want.

https://mycroft.ai/

Previously mentioned on Slashdot, Mycroft.ai can be built on a Raspberry Pi and perhaps other clones, and voice processing can be done locally. If I wanted something like this I'd probably use that.

Until these devices can do speech-to-text and home control and local resource interaction (your PC, basically) without going out on the net, there can be no reasonable expectation of privacy.

Closest so far is MyCroft. It's modular, it's open source, and so it has the potential to be as good as we want to make it.

connect it to the Google Assistant. Along with everything the Google Assistant already does, you can add your own question and answer pairs.

I'm investigating for myself this at the moment and I believe that the most agnostic one is currently Mycroft: https://mycroft.ai/about-mycro... but this still needs to be 'paired' with: https://home.mycroft.ai/. So it's a question of degree and who do you trust/want to support.

There's a niche for a full-stack open source one, I believe built from Sphinx etc.: http://cmusphinx.sourceforge.n... OK, I'm thinking like Stallman, but it's important not to get sucked into Google, Amazon and Facebook with the false lure of 'open source' NOT, as Wayne and Garth would say.

connect it to the Google Assistant. Along with everything the Google Assistant already does, you can add your own question and answer pairs.

I'm investigating for myself this at the moment and I believe that the most agnostic one is currently Mycroft: https://mycroft.ai/about-mycro... but this still needs to be 'paired' with: https://home.mycroft.ai/. So it's a question of degree and who do you trust/want to support.

There's a niche for a full-stack open source one, I believe built from Sphinx etc.: http://cmusphinx.sourceforge.n... OK, I'm thinking like Stallman, but it's important not to get sucked into Google, Amazon and Facebook with the false lure of 'open source' NOT, as Wayne and Garth would say.

Also https://mycroft.ai/ (from the comments on yesterday's story about FSF projects)

Free software assistant... already exists

http://mycroft.ai

They've got an RPi image you can download, slap on a card, and be up and running with a USB mic and something to handle the audio out.

Seems to me like the FSF should pay more attention to what is already going on.

The answer is MyCroft

I plan on buying one of these the very soonest I can once they are actually shipping the hardware. Echo is crippled by the many limitations Amazon coded in on purpose -- it's basically something that looks up text matches and does something if it finds one. No language parsing worth a damn. Even so, it's very useful, and within those limits, you can make stuff for it, Amazon's pretty open about it as long as you can set up a secure server (ugh) or use their cloud (double-ugh.) Siri, as per usual for Apple, is a much more closed system, and frankly, it's of no interest at all to me because of that.

Mycroft is completely open source. I have very high hopes for it because of that. I have reams of my own natural language processing code I should be able to plug right in the moment there is a speech-to-text engine I can use directly. Others do as well. Custom apps in the home space, that are actually somewhat smarter than...

[if string == "turn on light" then TurnOnLight]

I suggest everyone check MyCroft out. Perhaps you'll be as enthused as I. I can hope. ;)

Not yet mentioned yet is http://lucida.ai/ -- it's the successor to Sirius, and where all the ongoing development is focused.

Major options that are mentioned elsewhere in the thread:
https://mycroft.ai/ (One of the most advanced,can actually be used in a pretty useful manner now, but sends snippets to Google for voice recognition--they intend to change that eventually, and they don't have a full-time open mic. Plus they aggregate audio across users so it's less identifiable as from a single source).
https://wiki.mozilla.org/Vaani (from the Mozilla project; supposed to enter beta this month according to that page)

Thanks for asking the question. I didn't know about Mycroft until I looked for an Intelligent Personal Assistant.

This was the kickstarter: https://www.kickstarter.com/pr... Their main community website is: https://community.mycroft.ai/ They also have a slack here: https://mycroftai.slack.com/me...

Open API? Not good enough. I want to be able to have completely open Infrastructure under my direct control. Mycroft is the only AI that I'm probably going to build and rely on to operate my home. Why? Because if I'm going to have an assistant in my house, it's not going to be a closed box that someone else has the keys to with ownership over the hardware processing my information system. If I'm going to have an AI assistant, I'm going to be in sole control of the Hardware (on premisis), network infrastructure, and any APIs that operate the system. At least now I might have a use for my Tesla cluster that's been collecting dust (no I didn't pay nearly that much for the three of those).

So, that will be fine for you; but most people lack the time and/or the skills to make that a reality, even if they cared.

Open API? Not good enough. I want to be able to have completely open Infrastructure under my direct control. Mycroft is the only AI that I'm probably going to build and rely on to operate my home. Why? Because if I'm going to have an assistant in my house, it's not going to be a closed box that someone else has the keys to with ownership over the hardware processing my information system. If I'm going to have an AI assistant, I'm going to be in sole control of the Hardware (on premisis), network infrastructure, and any APIs that operate the system. At least now I might have a use for my Tesla cluster that's been collecting dust (no I didn't pay nearly that much for the three of those).

Mycroft just released (Python, version unspecified) code they say you can run on your Raspberry Pi; Mycroft is an Alexa-like system, differences being it's open, the s/w is free so you can build your own, and the hardware is pretty open too.

There is cloud STT (Speech-To-Text) going on, but they're interested in local STT according to an email they sent around to those of us on their mailing list. My GPS (ca. 2013) does non-cloud general STT, so there's working code out there.

Speaking as an owner of both an Echo and an Echo Dot, I'm very hopeful that Mycroft will join them, perhaps even replace them.

Echo's huge drawback is that it doesn't have a local operating mode via LAN ports, nor local STT and TTS. Not to mention the absurd requirement that you put up an SSL server just to make the simplest possible function work.

Just waiting for [...] a fully open source answer to this.

MyCroft may be of interest then.

I liked it well enough to get the Tap for the bedroom

We have an Echo (which we use quite a bit.... timers, alarms, news, weather, spelling, etc.) but for music.... the built-in speaker is low-fi and not satisfactory to me. So I bought an Echo Dot as soon as they came out, specifically because it had a line-out connection. That one, I use to listen to music here in my office (I have a very nice audio system in here), and I find the dot very satisfactory indeed in that role, although I do use the other features as well.

There do seem to be a lot of user-level haters; which leads me to believe there are a lot of people who've never actually used one. The claims that it has been a marketing failure are laughable; There is some approximate sales information, indications are that the product is doing quite well. Speaking as a user, I can understand why. We (my SO and I) find it very handy.

My objections to Echo / dot are about the developer ecosystem, the voice recognition implementation, and the secure server issues.

The "canned phrase" collection approach to command recognition is the very antithesis of any attempt to reach for "AI." It would have been wonderful if there was either a local interface so you could provide smarter processing, or if Amazon would actually provide smarter processing on their own. Between that, which is really a pretty crippling issue, and the requirement for an secure server with a +$ certificate for anything other than testing, Echo is not appealing to me as a development platform.

I've been watching MyCroft; that looks like it might have some potential.

Both presently suffer from online-only operation; the speech handling is dead if there is no connection. Hopefully that will be resolved in MyCroft's case, as it's actually an open system and they have mentioned that they're interested in pursuing local STT. My cheapo GPS ca. 2013 has reasonable general purpose offline speech recognition. However that was done, I would hope the underlying code would be better today, and I would love to see the capability in MyCroft, or Echo, or whatever. Being tethered to an active network is not a good thing.

Echo is fun to use. But the developer process... it's pretty much a grade-A clusterfuck.

To get Echo to do anything, you have to provide canned speech fragments that define everything your service is going to do. There's no "smarts" at all behind the speech interpretation; it's straight-up text-to-text matching. That's quite aside from the secure server requirements and the complete lack of a reasonable local echo-to-computer-to-echo interface.

I have an Echo and an Echo dot. Love using them. Wish they were reasonable to develop for, have many things I'd like to do right here. Looks like Mycroft is where I have to place my hopes, though. Amazon's not showing any signs at all of giving Echo anything more than the boneheaded API it currently has. Of course, it's designed to support Amazon. Not the customer. Which illuminates their approach somewhat, I think.

My GPS has speech processing. Works decently. Has to understand place names; and it does. It was an under $100 GPS. It does this without a connection to anything. And while it contains a huge map database. And it's not a very new2 device; I bought it in 2013 (and no doubt the speech recog software was designed somewhat earlier than that.)

The reason that they're all using the cloud isn't because they have to. It's because they can turn you into a product more easily; "you" being the developer and the user.

Best direction here is to dig out from under the commercial models and see to it that the field is actually open. Take a look at where MyCroft is aiming. They're shipping prototypes last I heard. They're looking at on-board speech recog as well.

Watch Mycroft. It has the potential to outdo Echo, and quite handily, too.

Looks a little dumb with the face and all, but it's OS, so you can hack (fix) that quite easily.

--fyngyrz
anon due to mod points

Regarding AI I'd take a look at the Mycroft project.

https://mycroft.ai/

Still getting started but have plans for AI on the desktop.