DeepSeek launched a free, open-source large language model in late December, claiming it was developed in just two months at a cost of under $6 million.
It’s also open source and you can run it locally without a super beefy GPU. The big deal is that it’s super light on resources compared to GPT, while producing comparable results. It’s only a question of time until someone makes a less censored version of it.
Either way, OpenAI is doomed unless they have something up their sleeves.
I am interested to see what the more knowledgeable folks come across when inspecting it and really testing it out. If it is legit, that is a huge step forward against these have-it-all tech bros blighting everything.
I don’t think they would’ve made it open source if they had anything to hide. Now there’s also many eyes on it, so we’ll probably learn soon enough if something is off.
Yup, that gives me hope. Also, I am aggressively anti-CCP, but very pro Chinese, so if the people that made it aren’t being manipulated by the CCP and some of the self-censor is just habit or survival, I get it. They have decent STEM education, if you are the right people, and they are an innovative people.
I’m not sure swapping out one set of tech bros for another set is changing all that much?
I think a lot of the costs of LLMs was always going to drive toward much lower numbers even given the software techniques more or less staying the same, given how the costs of electronics go…I already could run several models on hardware that I bought several years ago.
I do think it holds a lot of promise when properly applied, but yeah, a lot of the speculation is from the usual assholes hoping to “disrupt” things by throwing nearly everyone out of a job. I’m not sure these Chinese tech bros are going to be any more benevolent (in the long term) than the current set in the U.S. OpenAI started as a non-profit…
It’s also open source and you can run it locally without a super beefy GPU. The big deal is that it’s super light on resources compared to GPT, while producing comparable results. It’s only a question of time until someone makes a less censored version of it.
Either way, OpenAI is doomed unless they have something up their sleeves.
I am interested to see what the more knowledgeable folks come across when inspecting it and really testing it out. If it is legit, that is a huge step forward against these have-it-all tech bros blighting everything.
I don’t think they would’ve made it open source if they had anything to hide. Now there’s also many eyes on it, so we’ll probably learn soon enough if something is off.
Yup, that gives me hope. Also, I am aggressively anti-CCP, but very pro Chinese, so if the people that made it aren’t being manipulated by the CCP and some of the self-censor is just habit or survival, I get it. They have decent STEM education, if you are the right people, and they are an innovative people.
I’m not sure swapping out one set of tech bros for another set is changing all that much?
I think a lot of the costs of LLMs was always going to drive toward much lower numbers even given the software techniques more or less staying the same, given how the costs of electronics go…I already could run several models on hardware that I bought several years ago.
I do think it holds a lot of promise when properly applied, but yeah, a lot of the speculation is from the usual assholes hoping to “disrupt” things by throwing nearly everyone out of a job. I’m not sure these Chinese tech bros are going to be any more benevolent (in the long term) than the current set in the U.S. OpenAI started as a non-profit…