“In only about 5 years only, AI-driven smart assistants have become a common usage for many us. We find them on websites, to catch visitors attention or potential customers, we find them integrated in connected speakers, mostly by Google, Amazon and Apple, and of course they also are present in many iOS and Android smartphones. (…)”
This post is related to the announcement of the “Elivia” /e/ smart assistant initiative, and will welcome comments and suggestions.https://medium.com/@gael_duval/elivia-towards-an-e-smart-assistant-9fe30ccb28e1
Regain your privacy! Adopt /e/ the unGoogled mobile OS and online services
Hi @GaelDuval ,
It’s a very important subject, I’m happy that e foundation is working on it.
Few weeks ago, I read a french article 1 about a new virtual assistant with same goals, a project of Stanford University : OVAL : Open Virtual Assistant Lab 2.
It doesn’t seem that you evaluated it, but it seems it may be interesting.
Hope this helps.
This is an interesting development. I was wondering why MyCroft is not suitable?
I think the first step is a good STT engine: having the ability to use STT in the default keyboard would be a great improvement to /e/ OS. AI assistant is a plus, but I really think Speech to Text is a must.
Mozilla’s DeepSpeech work was promising, but might not be continued (https://www.phoronix.com/forums/forum/software/programming-compilers/1202341-mozilla-s-incredible-speech-to-text-engine-is-at-risk-following-layoffs), and tests made by Mycroft were showing that a lot of GPU power was still necessary to run it (https://mycroft.ai/blog/deepspeech-update/?cn-reloaded=1).
So… would be cool, but it might not be so simple to achieve…
I applaud the forward-thinking approach of this whole project, having an integrated AI assistant will definitely help making the whole ecosystem more attractive for people looking for alternatives to stock Android.
Definitely loving the idea of making it extensible too, through “skills” à la Alexa so that developers can create and share their own customized setups.
However I still think that the focus should be going towards launching a long-awaited version 1, stable and well functioning to avoid any onboarding disappointments and retain more users.
I actually think /e/ would be a great name for an AI! Though at this point seems unlikely /e/ will be rebranded to something different?
Ideed, this are valid issues. This is a complicated topic. Both STT and TTS are not on pair with closed source solutions. I guess this won’t change for open source without a sustained funding.