DeepSeek didn't release the code by accident. They know their code will be studied and incorporated into all the frontier models out there. They also know that USA has access to farms of 100s of thousands of GPUs. But they are not afraid. They have a game plan, that is certain.
It seems to me that the path to collective terrestrial intelligence, CTI, is the ability to merge, deduplicate, fact check, and aggregate the knowledge and sentiment expressed in public conversations with billions of people. How to make it open source, decentralized, and privacy protecting as possible?
Thank you! Starting 4:30 is pretty much the very question I've been looking to hear answered all week😊 mostly out of curiosity but it has huge implications for the development of future models. Chat GPT is impressive progress but it's still nothing now than a child's toy compared to what we will have in 10 years. Honestly, how quaint. So if they did it that way then there are multiple more thousand fold improvements to make before these models actually become kinda matured
So the US companies are upset because the info they stole was stolen by someone else. We americans should raise our taxes to help these rich people get richer. God Bless America
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
pretty rich for AI companies to complain about IP theft when their entire business is stealing other people's IP and plagiarizing it
DeepSeek didn't release the code by accident. They know their code will be studied and incorporated into all the frontier models out there. They also know that USA has access to farms of 100s of thousands of GPUs. But they are not afraid. They have a game plan, that is certain.
Did you listen to the podcast? They didn't release the code nor the training data. They only released the weights.
@@Nainara32They released the code. It is open source. That's how they know about all of these process improvements.
Plus the cyber attacks from the US 😂
Americans: We will win by using the gift you shared with us while restricting what you need to innovate further.
China has dozens of large languages generating AI programs, deepseek, Qwen from Alibaba, and several others.
It seems to me that the path to collective terrestrial intelligence, CTI, is the ability to merge, deduplicate, fact check, and aggregate the knowledge and sentiment expressed in public conversations with billions of people. How to make it open source, decentralized, and privacy protecting as possible?
Alababa just release another AI program Qwen, which is even better than deepseek.
Thank you! Starting 4:30 is pretty much the very question I've been looking to hear answered all week😊 mostly out of curiosity but it has huge implications for the development of future models. Chat GPT is impressive progress but it's still nothing now than a child's toy compared to what we will have in 10 years. Honestly, how quaint. So if they did it that way then there are multiple more thousand fold improvements to make before these models actually become kinda matured
monopoly is broken. Now game is on for everyone
So the US companies are upset because the info they stole was stolen by someone else. We americans should raise our taxes to help these rich people get richer. God Bless America