DeepSeek Mania Shakes AI Industry to Its Core
AI summary of the linked articleDeepSeek's release of its R1 model has sparked widespread alarm across the AI industry and financial markets. The Chinese-developed model reportedly outperforms OpenAI's and other companies' models on some benchmarks, despite being trained on older chips at a much lower cost, and it became the most popular app in the US Apple App Store. Nvidia's stock has tumbled, and Marc Andreessen called the model "AI's Sputnik moment." Because DeepSeek is free, modifiable and cheaper to run, experts are worried about the viability of OpenAI's already unprofitable subscription products. |
1 comment mishaSubmitter Fair enough, I suppose, but as the developers of “uncensored” AI models have been shouting from the rooftops since the beginning: any AI model that the user can’t control entirely is subject to censorship. OpenAI is a prude, and will refuse to engage users on a lot of topics, sometimes for reasons stated in OpenAI’s policy, and sometimes for reasons we’ll never understand because OpenAI is a black box. “Why should the open-source AI running on my computer, get to decide for itself when it wants to answer my question? This is about ownership and control. If I ask my model a question, I want an answer, I do not want it arguing with me,” Eric Hartford, a developer of uncensored AI models, told me last year. If anything, DeepSeek maps a better AI future for those concerned about censorship because it was released as an “open weights” model, meaning people could modify it to talk about Tiananmen Square and whatever else they want. |
More from 404media.co on Kiwi News
|