Hi everyone,

I just wanted to share a little bit of progress that we're making with Spencer's software. 

Here is a short demo of how it works and what it looks like right now. 

What we're currently working on is called intent detection or intent classification.

 

Intent detection is an integral part of all personal voice assistants or chatbots these days and it is basically a piece of code that makes the computer "understand" what you want. 

 

This is quite important because you can say that you want something in various different ways. 

For example, if you want your Spencer to tell you a joke, you can ask him to do it like this:

Spencer, please tell me a joke

or like this

I would like to hear something funny. 

 

Intent detection helps spencer understand that both sentences are a joke request and then he generates the answer and plays the sound file pretending that he is a human (but he is not, he is just a plastic box with a CPU inside).

We're smashing bugs and crunching the code as fast as we can. The schedule is tight, but we are doing our best to deliver this to you folks as soon as possible and before Christmas.

 

Thank you for being a part of our story.

- Albert and the team