Thanks. You should actually always harness your knowledge to advance other domains as everything is related, right? Good luck with your PhD studies.
@Salim

You're right, but keep in mind that it's equally hard to get VCs to invest in your vision. Especially that this not a SaaS and will obviously require a lot of funding and patience to get some results. But I will shoot my shot.

There are a couple of promising approaches out there that we looking to leverage such as Swin attention and other parallel data generation techniques, so we're going to build on that. One novel algorithm that explicitly aligns with our vision is tensor networks. These things aim to represent quantum states say entanglement in terms of diagrams and you can run them locally on classical computers. This is still an emerging field yet promising, so we definitely going to look closely at that.@Aya

You're right, that's we will clean up the dataset afterwards. But keep in mind that achieving as much bias as possible is intentional here. We want users to genuinely interact with the model in different life case scenarios so we can refine and train model to sophisticatedly output responses tailored to the user's mood, believes and temperament in general. It's similar to the nosing-denoising approach in stable diffusion, you have to understand the noise first to mitigate it and ensure the best outputs possible. @Ryner

Well, I believe that AGI is not the simulation of the human cognitive system but rather the stimulation of it. Generation of custom responses with respect to each context and user's behavior and attitude. Something fed and trained upon factual biases which will allow the machine to develop informed and legitimate opinions. Bias is inevitable and achieving an independent reasoning abilities on the machine level is impossible merely speaking. That's why this is critical and highly sensitive, and why we need to make sure that the ppl behind those biases are as transparent and authentic as possible. @MrWick

Could you briefly clarify why AI safety matters that much and that we don't even have the machine which is the source of all potential risks? All the best.

good luck with that

Some people were confused and were asking me if I'm looking to leverage the power of quantum computers warning me about how expensive they are to operate. I have to zero intention to do that. These quantum computers are very unstable and require extreme conditions to run. We believe that current compute is enough to test and validate our theories. the big setback is not resources but rather the vison, most people in the industry are pretty much trying the same thing in the same way, so we looking to break through this paradigm and try new things.@Eren

Entangled data is the premise of the core idea of compressing the data into low dimensional representations in order optimize the compute both qualitatively and quantitatively at the lowest costs possible. And yes, if we can achieve this, we can unlock a ton of possibilities and it all roots down to the data and its gestions, and little it has to do with architectures and training techniques.

@Jenex-