Preliminary Usage Evaluation of GPT-4o with Canvas

This model is the Beta version launched by OpenAI on October 3rd. It's a nice upgrade, but not revolutionary; we can see from this upgrade that the OpenAI Foundation is essentially still responsible for stock prices and shareholders. Canvas is a new interface that allows users to collaborate with ChatGPT for writing and code development, going beyond simple chat functions. With better real-time editing and adjustments (further promoting team collaboration); improved writing (supporting suggestions for editing, changing reading levels, adding symbols, etc.) & coding capabilities (supporting code audits, bug fixes, code porting, etc.) We can infer two general development paths for AI's future through Canvas: 1. Enhancing user experience based on large models 2. Empowering existing software or SaaS products with AI Overall, Canvas does not necessarily improve the quality of AI-generated content, but it can lower the barriers to using large models and enhance user experience. For those who are not very skilled in prompt techniques but have a deep understanding of the business, it can further reduce the difficulty of applying AI. I believe the significance of technological advancement—technological equality—has been achieved. However, for someone with average business capabilities and not much creativity or innovation, Canvas cannot help improve the quality of AI output. Therefore, the remaining rationality and innovative revolutionary power of humans compared to AI is still quite evident (at this stage). Finally, a side note: if Mr. Luo Ji, the wall-facer, sees OpenAI, perhaps he will no longer need Zhuang Yan, and humanity will lose a swordsman. After all, Luo Ji can solve his brainstorming and mental peaks just by consuming tokens.