Voice Capabilities Expansion

My name is Daniel Mueller. These past few months I have been navigating a company called Galacto Corporation. We made some findings in the area of voice enabled AI as it presents a wealth of opportunity to benefit those, especially with neurological disorders.

I ask if you utilize these concepts, you give some recognition, as this furthers our goal of educating and innovating in the field. We do not want to sequester this knowledge as we think it could be a vital aspect of the future of AI, but we have created working prototypes of our idea which could be optimized to grant a better functioning product.

It is much more comfortable to talk to the voice enabled version of GPT-4 on the iOS app. This yields better philosophical data. It is intriguing to bounce ideas off of it verbally, because it just feels much more human.

In conjunction with a search engine based context constructor you can recall remarkably large time-spanning entries of data to apply to the current context window. We did this in a rudimentary form with spaCy. It used direct keywords from the text to begin the search.

We found it is best applied when you leave it listening at all times, storing information privately for the user. The user may ask questions about real events which have occurred, as well as infer logical deductions. This is incredible for detailing a shopping list, summarizing meetings, and routine questions.

The speaking capabilities of Galacto are only activated when it’s name is said. This prevents it from responding constantly to things it should “ignore” at least in the moment.

We created Galacto to help a man with severe memory loss take control of his life. He has frequent memory loss events where he cannot recall how he got to where he is, who he is, or what he should be doing at any given moment. Galacto allows him to speak to someone that remembers for him.

This technology could change lives. We are pursuing funding routes but believe that this may be best left in the hands of those with the foundational models such as multilingual speech recognition.

Please do with this information what you think is best for the world. We are always here to chat.