Building an AI Agents WORKFORCE!!! [NEW PAPER] 

You made an absolute gem of a point. I listened to a podcast a few years ago that touched on this topic, How langauge defines us. Like for example some languages uses direction in such a way unlike others (almost similar to your point), I'll try to get the link and share here. Thank you for such a thoughtful comment and I'll pin this!

@@1littlecoder jazakallah khair. Yeah I would appreciate that .these tools are still black boxes. We need miners like you who go into these domains and mine new gems out of it . The more we will understand this tech the more we will be able to predict its consequences through its application. Coz currently we dnt understand much about them and so we are barely testing them .

Here is the podcast - www.npr.org/2018/01/29/581657754/lost-in-translation-the-power-of-language-to-shape-how-we-view-the-world (5 years ago)

Dear, it seems to me that you are reversing the causality relationship here. It's not the language that is intelligent (or contains elements of intelligence in itself); language only exists because there were rational and intelligent agents who were capable of creating it. Language serves a primarily social function and is nothing more than a means of organizing and communicating knowledge that already exists in people. The Russian words for all shades of colors will mean absolutely nothing to someone who hasn't learned Russian, just as the directions in African languages will have absolutely zero meaning to someone who is not familiar with them. You first need to adopt this language and have something before it for it to make sense. A child doesn't learn their mother tongue through the language itself; they learn it through the world and exploring the world, compartmentalizing the world in their 'internal world model,' which in turn will shape and be shaped by their language (and yes, language can 'shape how we see the world,' but we can shape language; otherwise, there would not be multiple languages or variations of them, and language itself would not change as it actually does). For a 'color' to make sense, you need to have had the experience of a color, and that's why it's not in the language itself, but in the complex interaction that starts in the agent and only then shapes and is shaped by it. As for the issue of language models, translation itself is an incredibly nuanced and complex task (see Quine on this subject), and these models being good at modeling natural language from natural language is, I believe, just a consequence of them being able to relate the context and position of words to the probabilities of responses, capturing the patterns that tens and/or hundreds of millions of humans had to put on the internet before that. The same technology that would allow for context-aware translations to be possible (and context is absolutely essential for a good translation, it's the difference between using word X and it meaning Y, Z, or W or completely changing the meaning of a sentence).

Thanks very much

Go for it! 🚀🚀🚀

Haha, That's quite possible

Yeah, with open interpreter@@1littlecoder

I guess saw this on Twitter. Why ?

@@1littlecoder Just curious. I was looking for sources to keep myself updated with the latest trends.