EP35 DeepSeek把AI价格打下去了,性能却比肩4o!TripleB 英语唠唠嗑

EP35 DeepSeek把AI价格打下去了,性能却比肩4o!

18分钟 ·
播放数473
·
评论数4

Summary:

We discussed DeepSeek-V3, a newly released open-source AI model from the Chinese company, Huanfang Quant. DeepSeek-V3 boasts performance comparable to GPT-4 and Claude, but at a significantly lower cost, achieved through efficient optimization of existing technologies rather than groundbreaking innovation. This cost-effectiveness is crucial for widespread AI adoption, compared to the historical development of the steam engine and railroad. Finally, emphasize the importance of China's role in driving AI development through cost reduction and performance tuning, presenting DeepSeek-V3 as a prime example of this approach.

Get the Doc 👉🏻 包含逐字稿、简介、时间线、FAQ、以及学习指引

Please let me know in the comments section if you have any queries.

展开Show Notes
老觉得我以后会学得像一个AI
JackyCc
:
其实不必有这种担忧,从物理层面人类老师和AI输出的都是音频信息传入你的耳朵,两者没本质区别。重要的是学语感、学用词、增加词汇量。你想学成AI也是不可能的,很明确的人类达不到AI的程度。学是输入,要消化,然后再输出,每个人输出的结果也会大不相同。AI仅仅只是工具,主动权在你手上!
姜总书J
姜总书J
2025.1.06
这个是真人吗?还是notebooklm做的?
JackyCc
:
notebooklm,我们做了中间层,优化过的