Skip to main content

Search CooLWJ

选择搜索范围，输入关键词后按回车。

Your keywords stay when you change scopeEsc to close

Language

Technology & everyday life

Development, AI tools, local voices and reading.

Write a post

AI & tools

1–5 / 5 articles

AI & tools

Testing Mainstream Large AI Models on Calculating Stock Returns

I used common Chinese and international AI models available on the market to help calculate my stock profit. The Chinese models included DeepSeek, Alibaba's Tongyi Qianwen, iFlytek Spark, ByteDance's Doubao, Moonshot AI's Kimi, Tencent Yuanbao, Baidu's ERNIE Bot, Zhipu Qingyan, Baichuan AI, 360 Zhinao, Tiangong AI, and SenseTime's SenseChat; the international models included GPT-4o, Claude-3-Opus, Gemini-2.0-Pro, and Grok. Given the same input, only DeepSeek calculated correctly.

837 reads

AI & tools

Model Distillation and Model Pruning

Model distillation and pruning are two important techniques for compressing neural network models. They can effectively reduce a model's parameter count and computational complexity, improving its operating efficiency and performance. Although both model distillation and model pruning are neural network compression techniques, their methods and objectives differ slightly. Model distillation focuses more on using a teacher model's predictive behavior to train a student model, while model pruning focuses more on removing unnecessary parameters and connections to compress the model.

759 reads

AI & tools

What Is a Vector Database?

A vector database is a collection of data stored in mathematical form. Vector databases make it easier for machine learning models to remember previous inputs, enabling machine learning to support use cases such as search, recommendations, and text generation. Data can be identified using similarity metrics rather than exact matches, allowing computer models to understand its context.

593 reads

AI & tools

After trying Kimi Explorer's AI search, Perplexity is left playing a supporting role

AI can answer questions, but can it truly “think”? Frequent users of AI search products will notice that these products often “drop the ball” when faced with complex questions—they can imitate Shakespeare's style and solve math problems, yet may be at a loss when tackling difficulties in everyday life. This is like a learned but inflexible bookworm, rich in knowledge but lacking flexible thinking. Clearly, simply turning AI search into a “talking search engine” is far from enough. So how can a large model with higher intelligence, stronger reasoning skills, and the ability to think deeply solve real problems for ordinary users, instead of merely solving Olympiad problems?

252 reads

Choose your theme

Your theme applies across the site.