Our Mission
We are going to change the world, making chatbots that actually comprehend what you and what they are saying. Even feelings will matter in this chat. It will have imagination and empathy. It will have a sense of urgency. And it will become your mate on searches for anything.
Our Story
Started in 2014 before the AI breakthrough of 2916 and the 2021 Chat GPT "explosion", Moshe Flam had a vision of technology which would constantly learn what is needed in order to teach itself what to do and then do it.
The most important point, says Moshe, an expert and perpetual student of brain and cognitive sciences is that it assumes it does NOT understand and goes about obtaining the understanding it needs, including your goals and requirements and its own responses.
The history of Overstand technology™ began in 2006 with a company called Intelligate (later acquired by AskMeNow) who came to me for technology assessment as the VP of R&D at MLL, at that time the 4th largest IT company in Israel. They had an ingeniously simple way of making a chatbot that could manipulate a predefined list of websites through conversation, bridging the gap between online chats in written natural language and online forms.
​
They had noticed that conversations allow information to be gathered slowly and through an interactive series of questions one could gradually gather the complete set of knowledge needed for any task.
​
At that time I was with ten years of research experience in the cognitive sciences at the IBM Scientific Center, with a focus on technologies of natural emergence, and a deep interest in human communication and linguistics. But the core of their technology was focused elsewhere.
I began looking into the possibility of achieving the same thing that Intelligate did, but in a completely different way, open to any website, and realized. that the key to everything would be an emerging web of meanings, dynamically bringing up clusters of meaning in every level and aspect of the conversations' analysis.
Every topic, each time and place of conversation, any field of discussion, all participants, every choice of words, all bring with them a complete set of meanings, in a chosen "dialect" with its language, lexicon, grammar, phrases, and extra-lingual nuances.
​
And that is what Overstand captures and overstands.
In 2018 I began programming AI chatbots and refreshed my knowledge about Deep Learning and Neural Networks.
​
In 2019 I went back to university, broadening my studies of philology, linguistics, sociology, and cognitive sciences.
Then, in 2022, I started competing with the new LLM technologies, studying them and developing with them and in them. Testing them and manipulating them. And, of course, competing with them.
The flaw in LLM technology is not a technical one, it is actually the basic unintelligent mistaken choice of goals:
​
You don't have to get it right.
You need to communicate what you may be getting wrong.
​
And then you need to listen.
Overstand technology gets that.
