Tag
Fun fact: new online side-job opportunities — find questions that Doubao answers incorrectly in biology/physics/math and get 400-600 yuan per item; find history questions Gemini gets wrong and get 600-800 yuan per item, reflecting the value of high-quality AI4Science samples.
Using the fastjson 0day to test LLM vulnerability discovery capabilities, found Claude and Doubao performed outstandingly.
The author reflects that the hospital is a good place to think about the value of AI, but the practical application of AI in healthcare has not yet brought universal breakthroughs; for example, AI tools like Fable 5 and Doubao have limited effectiveness.
The user shares their experience using the Doubao paid version to research AI products in Douyin videos, finding it functional but with low cost-effectiveness, fast quota consumption, and compares it with Codex Pro.
A long-term study involving 26,000 Chinese middle and high school students found that after students independently used AI, homework performance improved by 18%, but closed-book exam scores dropped by 20% within six months. Zhongkao and Gaokao scores dropped by 24% and 18% respectively, and 81% of students used AI to complete their homework.
This social media post shares a subjective observation: higher-income groups (A8 and above) tend to favor Doubao AI, whereas lower-income demographics are more likely to use programming-focused AI tools like Claude Code and CodeX.
ByteDance Seed insider Zhang Chi says the AI capability gap between China and the U.S. is widening, despite Doubao being China’s most-used chatbot.